ac200853aanalyzing nano material bio conjugates a review of current and emerging purification and...

Published: May 05, 2011

This article not subject to U.S. Copyright.Published 2011 by the American Chemical Society 4453 dx.doi.org/10.1021/ac200853a |Anal. Chem. 2011, 83, 4453–4488

REVIEW

pubs.acs.org/ac

Analyzing Nanomaterial Bioconjugates: A Review of Current andEmerging Purification and Characterization TechniquesKim E. Sapsford,*,† Katherine M. Tyner,‡ Benita J. Dair,§ Jeffrey R. Deschamps,|| and Igor L. Medintz*,||

†Division of Biology, Office of Science and Engineering Laboratories, Center for Devices and Radiological Health, ‡Division of DrugSafety Research, Office of Testing and Research, Office of Pharmaceutical Science Center for Drug Evaluation and Research,and §Division of Chemistry and Materials Science, Office of Science and Engineering Laboratories, Center for Devices and RadiologicalHealth, U.S. Food and Drug Administration, 10903 New Hampshire Avenue, Silver Spring, Maryland 20993, United States

)Center for Bio/Molecular Science and Engineering, Code 6900, U.S. Naval Research Laboratory, 4555 Overlook Avenue, S.W.Washington, DC 20375, United States

’CONTENTS

Complexities of the Nanomaterial-BiologicalConjugate 4454

Nanomaterial-Bioconjugation Chemistries 4455Properties of Nanomaterial-Biological Hybrids 4456Why Is Purification and Characterization

Important? 4457Purification 4458

Chromatography 4458Gravity Flow and Low-Pressure Liquid

Chromatography 4458

High-Performance Liquid Chromatography 4458Field Flow Fractionation 4458Electrophoresis 4459Centrifugation and Analytical Ultracentrifugation 4461Dialysis and Filtration 4462Extraction 4463

Characterization 4463Separation Techniques 4463

High-Performance Liquid Chromatography 4463Nanofluidic Size Exclusion 4463Field Flow Fractionation 4463Electrophoresis 4463Analytical Ultracentrifugation 4464Electrospray Differential Mobility Analysis

(ES-DMA) 4466

Scattering Techniques 4466Dynamic Light Scattering 4466Fluorescence Correlation Spectroscopy 4466Resonance Light Scattering Correlation

Spectroscopy 4468

Zeta (ζ) Potential 4468Raman Techniques 4470X-ray Diffraction 4470Small-Angle Scattering Techniques 4470

Microscopy 4470Electron Microscopy 4472

Atomic Force Microscopy 4472Near-Field Scanning Optical Microscopy 4472

Spectroscopic Techniques 4473UV�Visible Absorbance 4473Circular Dichroism 4474Fluorescence Spectroscopy 4474Infrared Spectroscopy 4475Nuclear Magnetic Resonance and Magnetic

Resonance Imaging 4475

Mass Spectroscopy 4476Thermal Techniques 4476

Modeling 4476Predictive Modeling 4479Structural Estimates 4479Protein Binding to Nanoparticles 4479Modeling of Quantum Dot Bioassemblies 4479

Emerging Technologies and Instrumentation 4481CytoViva with Hyperspectral Imaging 4481Xigo Acorn Area Analysis 4481Resonance Frequency Devices (Quartz Crystal

Microbalance and Suspended Cantilevers) 4481

Single Particle Tracking (Nanosight and Others) 4482Coulter Counter Devices 4482

Impact of NM Characterization on Nanotoxicology 4482Summary 4483

Author Information 4483Biographies 4484Acknowledgment 4484References 4484

The growing maturity of nanotechnology has led to theestablishment of more focused subdisciplines including

especially that of bionanotechnology. This field can more easily

Special Issue: Fundamental and Applied Reviews in AnalyticalChemistry

4454 dx.doi.org/10.1021/ac200853a |Anal. Chem. 2011, 83, 4453–4488

Analytical Chemistry REVIEW

be defined as the intersection of nanotechnology and biology andis loosely characterized by two somewhat converse foci: (i)exploiting nanomaterials (NMs) to investigate biological pro-cesses as exemplified by developing, for example, antibodyfunctionalized magnetic nanoparticles for in vivo tumor imaging,1

and in more simplistic terms, exploitation of a unique NMproperty in a biological setting and (ii) use of biological processesor systems to create, order, and investigate new nanoscalematerials or devices. A primary example of the latter would bethat of using DNA architecture and chemistry to control theplacement of nanoparticles in the pursuit of creating molecularelectronic devices at the sublithographic regime.2 NM-biologicalhybrid materials are considered “value-added” in that they arecapable of far more than each individual component alone. Eachparticipant imbues the composite with a unique property orfunction that is lacking in the other. For example, in (i) above, theantibody provides the biorecognition and targeting while thenanoparticle may allow for in vivo electromagnetic contrast. Thecomposite uniquely exploits each property to derive a novel,designer nanoscale entity. NM-bioconjugates have the potentialto revolutionize many fundamental and applied aspects ofscience and are already having significant impact in developingbiomedical therapeutics and diagnostics.1�6

It is important to begin by pointing out that what exactlyconstitutes a NM is still quite contentious and depends upon theperspective of origin. The initial working definition was anymaterial (biotic or abiotic, though not chemical) that was lessthan 100 nm in at least one dimension. Several national andinternational standards organizations have proposed acceptablenomenclature and terminology for use when describing nano-scale materials (see ASTM International E2456 and ISO techni-cal specification documents 27687 and 80004). While mostdescribe an upper size limit of approximately 100 nm in at leastone dimension, there is currently limited scientific evidence tostrictly support this value for all materials.7 This debate is nottrivial as it has important legal and regulatory implications.7Whatis not in dispute is that many of these materials display uniquenanoscale size-dependent properties that are of interest tobiology. These can include, for example, intrinsic properties suchas the quantum confinement displayed by semiconductor quan-tum dots,8 the paramagnetism of iron oxide and other metal alloynanoparticles,9 along with the conductivity and ballistic transportfound in nanoscale carbon allotropes.10 Beyond photophysicaland electronic properties, NMs, and nanoparticles in particular,have extremely high surface-to-volume (S/V) ratios (e.g., at<2 nm the S/V atomic ratio exceeds 50%), along with nontrivialsurface areas.4 This can allow for the display of multiple biologicalentities on their surface which can cumulatively provide in-creased avidity in certain configurations. NMs can also act as acarrier for an insoluble agent such as a drug. Alternatively, theNM can display multiple different biomolecules, thus imbuing thecomposite with multifunctionality. Indeed, one of the currentengineering goals is to create viable nanotheranostic agentswhere a NM would provide inherent fluorescent or magneticcontrast while displaying multiple targeting agents such as, forexample, tumor-specific antibodies along with cell penetratingpeptides and a chemotherapeutic agent. This nanomedicineconcept is being pursued to overcome many of the issuesassociated with current systematically delivered medicines.11

Some relevant discussions about the analytical challengesfacing researchers who wish to apply bionanotechnology,both from a physicochemical/surface property characterization

stand-point along with an understanding of the structure/func-tion and further interaction within biological systems are avail-able in the literature.4,12�16 Prior to a review of the techniquesavailable for purifying and characterizing NM-bioconjugates, abrief discussion of their structure, the chemistries used toassemble them, their physiochemical properties, and the drivingimpetus behind characterization is warranted.Complexities of the Nanomaterial-Biological Conjugate.

For the purposes of this review, the term nanomaterial (NM)will be used as an overarching term to describe materials,particles, or structures with a size dimension less than1000 nm. Where appropriate, the terms nanoparticle (NP),nanotube (NT), and nanorod (NR) and the specific constituenttype of material will be directly used in the text. We include NMsof all types and origin, be they metallic, semiconductor, alloys,oxides, etc. We further include multifunctional designer poly-mers and dendrimers, along with biotic NP-likematerials such asDNAOrigami, viral capsids, and large protein assemblies such aslight-harvesting complexes. These NMs either display novelnanoscale properties or when combined with other materialsprovide novel property or function. We inclusively definebiologicals as proteins, peptides, amino acids, nucleic acids,lipids, enzymatic cofactors, carbohydrates, drugs, and the like,and in certain circumstances, biocompatible molecules such aspolyethylene glycol (PEG).The complexities underlying the structure and function of

such hybrid materials are more easily illustrated by using NPs asan example. Composite NM and NP-biologicals can most simplybe described as multilayered structures, see Figure 1A. A givenNP sample, for example, is usually polydisperse with some finitedistribution of size. Further, its surfaces are not uniform butrather characterized by defects, edges, lattices, and vertices.4 Forbiological utility, many NPs consist of core or core-(shell)nstructures, where the outer shell(s) protect and insulate the corealong with mediating solubility and linkages to biomolecules, seeFigure 1. This concept is exemplified by the structure ofbiocompatible CdSe-ZnS core-shell semiconductor quantumdot (QDs), where the ZnS shell protects and passivates the corethus maintaining its optical properties while also preventingleaching.17 Most NPs are intrinsically hydrophobic and need tobe made hydrophilic and biocompatible by attaching a coating ofsurface ligands which act to mediate solubility or, more com-monly, colloidal stability. Ligand chemistry is complex and canrange from small charged molecules to amphiphilic block copo-lymers or dendritic structures which completely enwrap the NP.5

Ideally, the ligands would also provide chemical “handles” forattaching biomolecules of interest. Biomolecules usually make upthe outer layer of the composite structure, especially when theirintended function is for recognition or targeting. The complexityof these structures is further exacerbated by the sheer number ofdifferent structural iterations that are possible. This is highlightedin Figure 1A where eight different NP-biological assemblyconfigurations are schematically depicted. These range fromthe simplest, where the biological interacts with the NPcore directly while providing solubility, for example, to wherethe biological is entrapped inside the NP or the NP decorates thebiological. Figure 1B provides a slightly different, more-func-tional version of this composite structure. Here, the central NP issurrounded by a solubilizing PEG layer and then functionalizedwith a variety of representative biomolecules in three dimensions.In this example, the NP itself contributes to the desired utilitywith its intrinsic fluorescence.



Nanomaterial-Bioconjugation Chemistries.The number ofchemistries that have been applied to biofunctionalizing NMs isimmense and far beyond the current discussion.1 Despite thisdiversity, almost all can be grouped into three major categories:covalent chemistry, noncovalent interactions, and encapsulation.Figure 1C presents a schematic overview of the first twocategories using the interactions of a peptide with a NP.4 Forbrevity, encapsulating biologicals into NMs is not discussed here.The initial choice of which approach to utilize is usuallydetermined by a variety of factors such as the NM itself, alongwith its structure, size, and shape, the type of NM-surface ligand,and presence of available functional end groups, in combination

with the type of biomolecule, their size, available functionalgroups, chemical composition, and of course, what is ultimatelyrequired from the composites in practice. In the case of NPs, anexcellent discussion of how NP curvature, size, shape, and othercharacteristics influence the subsequent bioconjugation chemis-try is provided by Hamad-Schifferli.4

Covalent chemistries attach the biological directly to the NMsurface or to an intermediary which is most commonly a surface-attached stabilizing ligand.1,5 Bioconjugation to NM surfaceligands and other intermediaries, such as bifunctional cross-linkers, borrow heavily from standard bioconjugation chemis-tries; that is, chemistries developed to join/modify biomolecules

Figure 1. NM-bioconjugates: (A) Schematic of the various potential NP-bioconjugate components and configurations. Note: not to scale.(i) Biomolecule interacting with NP core, (ii) biomolecule interacting with a NP core via intermediate ligands, (iii) biomolecule interacting withNP shell layer that surrounds the NP core, (iv) biomolecule interacting with NP shell layer�NP core via intermediate ligands, (v) porous NP corecontaining entrapped biomolecules, (vi) porous or hollow NP core containing entrapped biomolecules surrounded by a NP shell layer, (vii) NP core(or NP core/NP shell structures) particles smaller in size than the much larger biomolecule, (viii) NP core (or NP core/NP shell structures) particlessmaller in size than the much larger biomolecule attached via intermediate ligands. (B) Multifunctional NP assembly: A representative NP decoratedwith multiple disparate functional molecules (e.g., nucleic acids, proteins, drugs, peptides) is depicted. Although shown as a flat representation, theconstruct would display biologicals in three dimensions. Robust conjugation of biomolecules to the NP surface is critical for the development of such“value-added” constructs that can provide multiple functions within one active NP platform. (C) The four general schemes routinely used for theconjugation of peptides to NP materials: (i) electrostatic interaction, opposite charges on the surface of the NP and the peptide are used to mediatecharge�charge-based NP-peptide assembly; (ii) direct interaction, certain peptide motifs can bind to/coordinate with the NP surface with high affinity,and examples include the interaction of free thiols with the surface of AuNPs and the high-affinity coordination of polyhistidine tracts with NPs(e.g., QDs) displaying Zn2þ-bearing surfaces; (iii) secondary interactions, this scheme utilizes specific ligand-receptor interactions and is exemplified bythe biotin-streptavidin (SA) interacting pair. The incorporation of the biotin moiety at the peptide’s terminus canmediate direct assembly of the peptidewith the NP. (iv) Covalent chemical attachment, these linkages utilize classical bioconjugation chemistry such as EDC-based coupling of amines tocarboxyls or NHS- and maleimide-mediated conjugation to amines and thiols. The figure concept adapted with permission from ref 4 and IOPPublishing Ltd. Copyright 2008 IOP Publishing Ltd.

http://pubs.acs.org/action/showImage?doi=10.1021/ac200853a&iName=master.img-000.jpg&w=475&h=351



with dyes, haptens, and other linkers. Hermanson’s BioconjugateTechniques continues to be one of the best resources on thissubject.18 Because of the exquisite specificity that is inherentlyafforded, recent research has focused heavily on adapting che-moselective or bioorthogonal chemistries such as the superfamilyof “click” reactions to modifying NMs and NPs in particular.19

Noncovalent attachments can be subgrouped into those basedon electrostatic interactions along with those based on biologicalrecognition/binding, affinity interactions, or enzyme activity.Electrostatic interactions require that both materials displayappropriate cognate surfaces to mediate attachment. Somebiologicals such as DNA have inherently strong negative chargeswhich can facilitate such an interaction. Biologically drivenbinding strategies are typified by biotin�avidin binding andagain require the requisite display of each participant in theappropriate configurations. The ability to synthesize peptidesand nucleic acids in a biotinylated form along with access to amyriad of reagents allowing site-specific biotinylation and the strongintrinsic affinity of this reaction (Ka∼1015 M�1) have made thisone of the most popular NM-bioconjugation approaches.18

Direct surface interactions are usually “dative” bonds as exem-plified by gold-thiol chemisorption. Thiolated peptides, proteins,nucleic acids, etc are commonly attached to gold (or other noblemetal) surfaces or NMs using this approach.20 Another examplewould be metal affinity coordination between polyhistidinesequences and the ZnS on semiconductor nanocrystalsurfaces.8,17 A variety of enzyme based approaches are alsoavailable where a NM surface is premodified with a targetmolecule and then exposed to a chimeric enzyme-protein fusionconstruct that binds the target. An example of this strategy wouldbe that of glutathione-S transferase fusion proteins binding toglutathione-modified surfaces.21

Considering the nature of all these chemistries, it is critical tokeep in mind that NM-bioconjugation typically occurs via astochastic process resulting in a distribution of NMs functiona-lized with different numbers or populations of biomolecules, seeFigure 2.22 Conjugates may consist of NMs both with and

without attached biomolecule or alternatively a narrow or broaddistribution in the ratio or valence of attached biomolecule.Depending upon the assembly chemistry utilized, there may alsobe heterogeneity in the orientation of the attached biomolecules.This is especially true where the chemistry reacts with multiplegroups such as when implementing carbodiimide (EDC) chem-istry to form an amide bond between the ubiquitous carboxyls oramines present on proteins and the cognate target group presenton a NM.18 These same types of reactions also have a highpropensity for cross-linking and forming higher order structuresand aggregates. Similar issues can be expected for biomoleculesdisplaying multiple biotin groups when presented to NM-avidinconjugates, given the latter’s multiple binding sites.Properties of Nanomaterial-Biological Hybrids. With the

characterization of NM-bioconjugates, a number of physico-chemical metrics are of particular interest: including NM sizeand size distribution, shape and aspect ratio, aggregation/agglomeration state, purity, chemical composition, surface char-acteristics, ζ (zeta) potential (overall charge), surface area,stability, and solubility, see Figure 2.15,23,24 Bioconjugation,however, adds additional questions and metrics to this equationincluding (1) confirmation of and type of biomolecule attach-ment, (2) average ratio of NM/biomolecule and ratio distribu-tion, (3) hydrodynamic radius, (4) structure and orientation ofthe biomolecule upon attachment, (5) separation distancebetween the biomolecule and the NM, (6) stability of thematerial during conjugation and of the resulting compositewithin the NM environment during the intended application,and (7) activity of the biomolecule upon attachment.4 In thecontext of attaching an antibody to a given NM (Figure 1B), forexample, it would be important to know the ratio attached perNM along with the antibody orientation and thus the availabilityof the active binding sites. Depending upon the attachmentchemistry used, the NM conjugates may also be cross-linked andcontain aggregates of various sizes. These descriptive metrics becomeexponentially more complex when examining NMs functionalizedwith multiple (different) biological and chemical entities.

Figure 2. Schematic highlighting issues pertinent to the purification and characterization of NP-bioconjugates.




Why Is Purification and Characterization Important? Sim-ply put, given just some of the relevant issues listed above,rigorous attention to purification and characterization methodol-ogies during the design and manufacture of NM-biocompositesare essential to achieving reproducible and well-controlledperformance in the intended application, see Figure 2. Thisconcept was eloquently iterated by Royce Murray, the editor ofthis journal, in a recent perspective on this subject “As importantas these applications are or may become, researchers sometimescharge into NP use with inadequate attention to what the NPsactually are. When used as chemical substances, or carriersthereof, NPs should not be deployed in ignorance of theircomposition and, ultimately, structure”.25 Adequate character-ization is also fundamental in the emerging field of nanotoxicol-ogy (discussed below).13,23,24,26,27 Furthermore, it will not belong before such comprehensive characterization of NM-bio-composites will be essential and even mandated in the research,

commerce, and regulatory sectors, to ensure reproducibility ofsynthesis, quality assurance, predictable behavior in intended use,along with safety and effectiveness.28,29 A somewhat simpleranswer to the above question comes in the form of a pragmaticquestion straight from the bench researcher’s perspective,namely, “What have I made and how do I confirm this prior tothe final application?”There are already in place a variety of established analytical

methods that have been successfully applied to the characteriza-tion of bare and modified NMs.4,30�34 These are well-developedtechniques that primarily grew out of the colloidal chemistry fieldwhich, for all intents and purposes, has now “morphed” into theNP chemistry field. Beyond a cursory introduction, we do notreview the characterization techniques themselves here butrather focus on their applicability in the current context. Theaim is to provide an updated overview of current and emergingtechnologies that have particular utility in the characterization of

Figure 3. Chromatography and field flow fractionation. (A) Chromatography: (i) schematic demonstrating the affinity chromatography-basedpurification of QD-antibody bioconjugates from unbound antibody. The QDs are prelabeled withMBP prior to antibody exposure. The antibodies werebiotinylated and bound to avidin electrostatically attached to the QDs. With the use of the affinity of MBP for the amylose column, the QD-antibodybioconjugate binds to the column while unbound antibody is washed away (step 1). The QD-antibody bioconjugate is eluted from the column viaaddition of a maltose containing mobile phase that competes with the amylose stationary phase for the MBP binding site causing displacement (step 2)The QD-antibody immunoreagent is then directly applied to the immunoassay.36 (ii) HPLC purification of 70-base polyT DNA conjugated to 5 nmAuNPs. When the sample was purified by HPLC, multiple peaks were observed corresponding to AuNP with differing stoichiometries of the attachedDNA. Figure reprinted from ref 42. Copyright 2008 American Chemical Society. (B) Field flow fractionation (FFF): (i) schematic diagram of anasymmetric-flow field flow fractionation (AF4) channel illustrating the modes of flow and principle of hydrodynamic separation. Reprinted withpermission from ref 50. Copyright 2010 Springer ScienceþBusiness Media. (ii) AF4 fractograms of chitosan-rhodamine and DNA/chitosan-rhodaminemonitored by UV�vis showing normalized absorbance at 260 nm as a function of the elution time. Reprinted from ref 51. Copyright 2010 AmericanChemical Society.




NM-bioconjugates.30 In many cases, it will be apparent thatestablished techniques are being applied and/or interpreted innew ways. Given the scope and breadth of this field, rather thanbeing all inclusive, selected characterization examples are high-lighted from the recent literature along with a critical discussionof the benefits and liabilities of each technique where warranted.We extend our apologies for the many omissions. As there are noreviews in this subject area that precede this one, we includereferences that extend beyond the last 2 years where pertinent.Sincemany of the techniques utilized here have found dual-use inboth purification and characterization of NM-bioconjugates, webegin with an overview of the relevant methods of purification.

’PURIFICATION

In an ideal world, following bioconjugation, purification wouldbe capable of separating the NM-bioconjugate from both un-conjugated-biomolecules and unmodifiedNMs while also havingthe ability to resolve discrete subpopulations of the NM-biocon-jugates, see Figure 2.4 While certain NMs, such as those that aremagnetic, provide inherent properties that make purificationrelatively simple (i.e., collection using a permanent magnet),35

most rely on several commonly utilized methods to minimallyseparate the unbound biomolecule from the NM-bioconjugate.Chromatography.These techniques separate based upon the

differing affinities of the multiple sample components for thechosen chromatographic mobile- and/or stationary/solid-phase.The actual separation mechanism depends on the chromato-graphic method utility but, in all cases, care should be taken tolimit nonspecific interactions with the stationary phase ordenaturation due to solvents which can be problematic forsensitive or labile biomolecules.Gravity Flow and Low-Pressure Liquid Chromatography.

Ion-exchange, size exclusion, and affinity chromatography, rununder either gravity or low-pressure conditions, are commonlyapplied to NM-bioconjugate purification and can even beextended to materials of quite large mass.36�40 Sapsford andco-workers, for example, used ion-exchange chromatography toseparate dye-labeled cowpea mosaic viral particles (CPMV, MW5.6 � 106 kDa) chemically functionalized with antibodies fromunbound antibody (MW ∼150 kDa) based upon their differingaffinities for a Q-10 Sepharose column under a salt gradient.37

Increasing the NaCl concentration altered protein�media inter-actions allowing the CPMV-antibody bioconjugate (elution at∼0.5 M NaCl) to be collected after the free antibody (elution at∼0.2 MNaCl). The colabeled dye/antibody viral nanoplatformswere subsequently applied in immunoassays and demonstratedimproved limits of detection (LOD) compared to dye-labeledantibodies alone. Goldman et al. used an affinity chromatographytechnique to separate antibody-labeled QDs from the unboundantibody. For this procedure, QDswere noncovalently decoratedwith both maltose binding protein (MBP) and antibodies. MBPwas recombinantly engineered to express a positively chargedleucine-zipper domain which electrostatically interacted with theQDs-negatively charged surface ligands. The antibodies werebiotinylated for attachment to avidin which was also electro-statically coupled to the QD surface. MBP’s natural affinity foramylose media bound and immobilized the QD-protein complexin the column where unbound antibody could be easily washedaway, see Figure 3A.36 Antibody-QD conjugates were then elutedfrom the column via the addition of maltose-supplementedbuffer and subsequently used for detection in four color

multiplex toxin immunoassays. In this example, the elutedconjugates were used directly; however, depending upon appli-cation, and excess salt or other eluent present such as maltosemay need to be removed first. Affinity-based methods do havethe additional drawback in that they require theNM to be labeledwith an affinity tag (e.g., biotin, His-tag, MBP, peptide). This canincrease the complexity of the resulting bioconjugate, along withthe requisite synthesis and purification steps and may impact theintended application as the NM surface must now be partiallyallocated to accommodating and displaying another potentiallyinterfering or complicating species with a different intendedutility.High-Performance Liquid Chromatography. High-perfor-

mance liquid chromatography (HPLC)41 coupled with reverse-phase,22 ion-exchange,42 and size exclusion (SEC)34,41,43,44

stationary phases have been extensively used for purification ofNM-bioconjugate from excess biomolecules. Optimized HPLChas demonstrated the ability to resolve discrete NM-bioconju-gates each displaying a different NM-to-biomolecule ratio.22,42,45

Alivisatos’s group elegantly demonstrated this ability usinganion-exchange HPLC combined with increasing salt concentra-tion in the mobile phase to elute and purify gold NPs modifiedwith 0, 1, 2, or 3 DNA molecules for subsequent assembly intoplasmonic structures, see Figure 3A.42 They also found thatelution time was dependent on the length of the DNA sequenceconjugated to the gold NP. Importantly, the high pressure andbinding/elution from the column did not compromise the NP-bioconjugate structure or functional integrity. In contrast tomoststandard HPLC applications, elution with high concentrationsof organics (i.e., acetonitrile) and inclusion of strong acids(i.e., trifluoroacetic acid) need to be carefully considered whenpurifying NM-protein conjugates given the possibility ofdenaturation.Stravis and co-workers recently demonstrated a nanofluidic

size exclusion technology capable of separating and characteriz-ing the size of 100 and 210 nm diameter fluorescent NPs withstrong potential for purification and size characterization of NM-bioconjugates.46 The benefit of this approach would be thatsample volumes can be significantly reduced, with the samplereservoir holding as little as 10 μL. Overall, chromatographytechniques are growing in popularity for a number of interrelatedreasons. These include relative cost, ease of use, wide range ofcommercially available instruments and column materials, andfacile tailoring of a given procedure for a specific NM-bioconju-gate system. Moreover, as these are well-established techniques,the requisite equipment may already be available in a particularfacility.Field Flow Fractionation. Field flow fractionation (FFF)

encompasses a family of analytical techniques in which thesample is introduced into a pressure driven mobile phasecontained within an open channel (no stationary phase) demon-strating a parabolic flow profile, and an alternate field is appliedperpendicular to the direction of flow.47 Typically appliedfields include cross-flow (direct or asymmetric), centrifugal(sedimentation), electrical fields (charge), thermal/temperaturegradients, magnetic, and dielectrophoretic fields. The principlesof separation are dependent on the applied field but thetechnique has demonstrated the ability to purify NM-bioconju-gates from both unmodified NM and free biomolecules, asrecently reviewed.30,48,49 Preoptimization of the buffer type,ionic strength and membrane used is required to obtain thedesired FFF separation and limit any nonspecific binding that can



occur at the accumulation wall.50 In a demonstrative example,Maand co-workers used asymmetric flow-FFF to separate DNA/chitosan nanocomposites from chitosan, see Figure 3B.51 Thesestructures were assembled to investigate their gene therapypotential, and the customized instrument allowed simultaneousUV�visible, multiangle light scattering (MALS) and dynamiclight scattering (DLS) detection, which along with purification,provided insight into the amount of unbound polycation, hydro-dynamic size, and size distribution. Commercial FFF instrumentshave recently become available which should aid in the wideradoption of this technology.Electrophoresis. Slab/plate gel- and capillary-electrophoresis

(CE) are the two principle types of electrophoretic techniques

commonly applied to the purification of NMs and NM-bioconjugates.52,53 Gel electrophoresis monitors the electro-phoretic mobility of charged species in a gel matrix, typicallyagarose or polyacrylamide, when an electric field is applied acrossit. For both NMs and NM-bioconjugates, the overall size, shape,and charge density influences the direction and distance movedin the gel.52,53 On the small scale, these techniques are routinelyused to separate and purify NM-bioconjugates and are also quiteoften used as a rapid and powerful tool for confirming biomo-lecular attachment to the NM scaffold through discrete changesin mobility, see Figure 4A,B.34,42,52�61 Separated NM-samplebands can also be extracted from the gels for subsequentapplication or further characterized using techniques such as

Figure 4. Gel electrophoresis and extraction techniques for purification. (A) (i) Gel electrophoresis purification of AuNP pyramidal structures, next toincomplete structures as standards. The direction of migration is bottom to top: (ii) (1) bare gold, (2) monoconjugate, (3�5) all possible two-strandproducts, (6a) pyramids formed by mixing all strands at once, and (6b) pyramids formed by mixing strands in pairs and then combining the two-strandproducts. No difference in yield was observed with this change in protocol. (iii) Typical TEM images of Au DNA-nanocrystal pyramids made from fourstrands of DNA. Images reprinted with permission from ref 62. Copyright 2009 American Chemical Society. (B) Separation of QD-ND/EB1 complexesby 0.5% agarose gel electrophoresis: (i) QD-ND loaded withNi2þwas incubated with increasing concentrations of His-tag EB1 protein (n = ratio of EB1to QD calculated fromUV absorbance) for 1 h prior to loading and running the gel, (ii) model of QD/EB1 interacting with a microtubule, (iii) temporalimage sequence (5 s/frame) of a single QD/EB1 moving on a microtubule (see arrow). Arrows in parts A (i) and B (i) indicate the direction of NPmigration. Reprinted from ref 63. Copyright 2009 American Chemical Society. (C) Chemical extraction: (i) chemistry used by Zhang and co-workers tofunctionalize QDs with amino acids, (ii) transition of QDs from the organic (bottom) to aqueous phase (top) following the DTC-Lys reaction andligand exchange. Reprinted from ref 83. Copyright 2010 American Chemical Society.




Table1.

Separation

Techn

iquesa

technique/types

NM-bioconjugates

analyzed

advantages

disadvantages

refs

chromatography

(typicalstationary

phases

include;

reversephase,ion-exchange

and

size

exclusion.)

1.standard

liquid

chromatography

2.high-perform

ance

liquid

chromatography(H

PLC)

3.hydrodynam

icchromatography

AuN

P-DNA,A

uNP-cytochromec,

QD-antibody,QD-PEG

1.used

topurifyNM-bioconjugates.

2.high

resolvingpower,insomeinstancescan

resolvedifferentN

M-to-biom

oleculeratio

s.

3.provideinform

ationon

distrib

utions.

4.canbe

used

toinvestigatepostproductio

n

stabilityandimpurities.

5.simpleandcost-effectiveto

use.

1.nonspecificinteractions

with

the

stationary

phase.

2.requiresoptim

izationfortheparticular

system

underinvestigation.

22,42,45,46

fieldflow

fractio

natio

n(FFF

)

1.sedimentary

2.electrical

3.flow

4.thermal

5.magnetic

6.dielectrophoretic

QD-DNA,polym

erNP-peptides,

polymerNP-drug

1.sedimentary

andflow

canprovideeffectivemass,

hydrodynam

icradius,densityandvolume.

2.provides

size

andsize

distrib

ution.

3.used

topurifyNM-bioconjugates.

4.thermalFF

Fcanseparatebasedupon

size

and

surfacepotential.

1.nonspecificinteractions

with

the

accumulationwallcan

beproblematic.

2.requiresoptim

izationof

theseparatio

n

conditionsfortheparticularsystem

under

investigation.

48�5

1,84�8

6

electrophoresis

1.slab

gel(includingagarose,

PAGEb -nativeand

SDS-PA

GEb-denatured

2.capillary

(CE)

AuN

P-DNA,Q

D-m

altose

binding

protein(M

BP),Q

D-BSA

,silicon

NP-streptavidin,ironoxideNP-

protein/antibody,PE

Gpolymer

NP-protein,polymerNP-drug

1.canseparateandpurifyNM�b

ioconjugates.

2.canprovidehydrodynam

icradius,bothrelative

andabsolute,and

ζpotential.

3.separatedmaterialscanbe

extractedand

characterized

byadditio

naltechniques.

4.separatio

ncanbe

dependento

nNM

shape.

5.simpleandcost-effectiveto

use.

1.forCEpotentialnonspecificinteraction

with

thecapillary

wallcan

beproblematic.

2.effectiveCErequiresextensive

optim

izationpriorto

analysis.

3.sm

allscaleuse.

4.requirescalibratio

nwith

know

nsize

standardsforabsolutevalues.

52�5

8,64,70,89,90

centrifugation

1.centrifugation

2.ultracentrifugatio

n

3.gradient

centrifugation

4.analyticalultracentrifugatio

n

(AUC)

a.sedimentatio

nvelocity

b.sedimentatio

nequilibriu

m

layereddoublehydroxides

(LDH)-

enalaprilate,Au-streptavidin,A

u-

adenoviru

s,QD-BSA

,QD-

dihydrolipoicacid(D

HLA

),QD-

PEG,silica

NP-protein

1.rapid,easilyadaptedto

differentm

aterials,cheap

andsimpleto

use.

2.AUC:size,size

distrib

ution,andshapecanbe

determ

ined.

3.AUC:structuraland

conformationalinformation

abouttheconjugated

biom

oleculecanbe

determ

ined.

4.AUC:N

Mmolecularweightcanbe

measured.

5.AUC:self-associatio

n/aggregationandother

interactions

canbe

investigated.

6.AUC:p

urificatio

nandstoichiometry

canbe

determ

ined.

7.sm

allsam

plesize

requiredandtechniqueis

nondestructive,hencesamplecanbe

extractedand

furthercharacterized

usingadditio

naltechniques.

1.AUC:expensive

andnotasintuitive

as

simplecentrifugation.

2.canbe

difficultto

recoverseparatedmaterials.

3.notquantitative(exceptfor

AUC).

4.cancauseadditio

nalself-associatio

n

ofthematerials,w

hich

may

cause

misinterpretatio

nof

separatio

nprofiles.

77,91�

95

aGeneralinform

ation:

Typicallycost-effectiveandrelativelysimpleto

use.Routin

elyused

topurifynanomaterial(NM)-bioconjugates.Can

provideapproximatehydrodynam

icradius,purity

ofproduct,

NM-to-biom

oleculeratio

,postproductiondegradation,andimpurities.

bPA

GE,

polyacrylamidegelelectrophoresis;S

DS,sodium

dodecylsulfate.



AFM, mass spectrometry, etc.34,58,62 For example, B€ucking andco-workers made CdSe/ZnS and InP/ZnS QDs water-solublethrough various ligand exchange methods and then attachedbovine serum albumin (BSA, used as a model protein) throughnoncovalent interactions to increase long-term stability. Agarosegel electrophoresis was used to separate the resulting BSA-QDconjugates from excess BSA present in the reaction mixture. TheBSA-QD product was then extracted from agarose gel slices bysoaking overnight in buffer.58 Alivisatos’s group used a similarstrategy to purify unique pyramidal nanostructure assembliesprepared from various sizes of gold nanocrystals functionalizedwith complementary single strandedDNA, see Figure 4A.62 Herethe DNA-nanocrystal pyramidal-bands were either cut out of theagarose gel and extracted into buffer or the gel was cut ahead ofthe band, filter paper inserted, and the gel run an additional 10min to push the band onto the paper. Once purified, transmissionelectron microscopy (TEM) was performed to verify the nanos-tructures and characterize their novel chiral properties whicharose from their tetrahedral symmetry, see Figure 4A.In another elegant demonstration, Dif and co-workers synthe-

sized QDs displaying mixed peptidic-PEGylated surfaces to bindpolyhistidine-tagged (His)n proteins.63 Gel electrophoresis in0.5% agarose was then applied to demonstrate successful QDbinding of a recombinant (His)6-tagged microtubular associatedend binding protein-1 (EB1) followed by purification of the QD-conjugate displaying the desired 1:1 stoichiometry, seeFigure 4B. Clearly as the ratio of EB1 to QD increases in theimage, the corresponding QD-EB1 bioconjugate migration ratein the gel decreases in a proportional manner. Distinct anddiscrete bands were visible at the lower ratios which allowed theauthors to purify the 1:1 conjugate. This was important as the gelconfirms that the 1:1 assembly is only a minor species in thatparticular reaction mixture with most of the QD remainingunconjugated; if left unpurified, the brightness from theseunconjugated QDs would significantly complicate the resultingexperiment by interfering with the ability to track single QDs.Such a low concentration of protein-conjugated QD in this

reaction is to be expected due to the underlying Poissonianassembly kinetics when utilizing this type of assembly chemistrywith the intention of creating these two particular conjugateassemblies, namely, those displaying one-protein and thosedisplaying none.64 Indeed, this gel result is also an excellentreminder of the stochastic nature of NP-bioconjugation and howreaction stoichiometry influences the resulting conjugate. Mon-ovalent QD-EB1 bioconjugates were successfully used to moni-tor the interaction of EB1 with microtubules during mitoticspindle formation in Xenopus cell extracts, see Figure 4B. Asevidenced in these representative examples, gel electrophoresisclearly has the ability to separate discrete ratios of NM-biocon-jugates; however, this technique can be limited by the complexinterplay of biomolecular size and chemistry, NM size, NMsurface character, along with gel sieving and separation capacity.In other words, not every type of NM-bioconjugate will beamenable to monovalent resolution using this approach. Tospecifically address these types of resolution issues, Liu and Gaorecently engineered a hybrid polyacrylamide-agarose gel systemwhich demonstrated the capacity to separate monovalent anti-body-functionalized QDs from a conjugation mixture.65 Com-mercial QDs solubilized with an amphiphilic polymer and surfacelabeled with multiple SA moieties were exposed to antibodiesthat had been reduced and biotinylated on the hinge regionsulfhydryl groups. The reactions were then separated in the

hybrid gel media and the monovalent QD-antibody conjugateextracted for subsequent use in Western blotting and cellularimaging where they demonstrated improved quantitative abilities.In its first or “classical” iteration, CE measured the electro-

phoretic mobility of charged species in an open capillary (nosolid matrix) filled with a liquid electrolyte under an electric field.Through a combination of sample component electrophoresisand electrolyte buffer electroosmotic flow (EOF), sample com-ponents are transported from the positive anode to the cathodewhile separating in a manner that is based on size-to-chargeratio.52,53 UV�visible and fluorescence spectroscopy are typi-cally coupled with CE for detection.52 Extensive optimization ofthe technique is almost always required for effective CE, asoutlined in a recent review by Surugau and Urban.52 The abilityof CE to purify NM-bioconjugates from free NMs has beendemonstrated for QD-BSA66 and silicon-NP SA conjugates.67

Variations on traditional CE include capillary gel electrophoresis,micellar and microemulsion electrokinetic chromatographies,68

capillary isoelectric focusing, and capillary isotachophoresis,although the later have been less commonly applied to NMs.69

Other electrophoretic methods with strong potential for applica-tion to NM purification include dielectrophoresis,52,70 isotacho-phoresis,58 and isodielectric separation.71 One of the benefits ofCE is its inherent high sensitivity, which allows for very smallsample sizes/volumes to be used. This, however, suggests that asmost commonly implemented, this technique may be moreappropriate for analysis and characterization rather than bulkpurification.Centrifugation and Analytical Ultracentrifugation. Centri-

fugation is a relatively simple and cheap technique that can be usedto purify functionalized NM-bioconjugates from unconjugatedbiomolecule. Depending on the density, size, and structure of theNM and biomolecules, application can sometimes be as straight-forward as using a benchtop centrifuge to separate functionalizedNMs from the reaction mixture and then either removing thesoluble bioconjugate or, alternatively, resuspending the bioconju-gate precipitate in the buffer of choice. For example, such amethodhas been used to purify layered double hydroxide NPs intercalatedwith the acetyl choline esterase (ACE) inhibitor enalaprilate fromfree drug.72 For smaller NMs, or those with a lower density,ultracentrifugation may also be useful for separation and purifica-tion. This is often used in gold NP-conjugate purification, forexample, with gold-NP-SA conjugates,73 and has also been appliedto isolation of dye- or PEG-modified potato virus X.61 In otherinstances, glycerol, sucrose, salt (cesium chloride CsCl), or othergradient-based methods can be used to help separate out NMconjugates from unconjugated materials during ultracentrifuga-tion. For example, Saini and co-workers genetically modified thehexon protein of adenovirus NPs with a (His)6-affinity tag thatallowed subsequent noncovalent coupling to 1.8 nm gold NPs(AuNPs) modified with nickel(II) nitrilotriacetic acid (Ni-NTA)to the virus surface.74 A CsCl centrifuge gradient was then used toseparate gold NPs conjugated with adenovirus from unconjugatedAuNPs and adenovirus.74 The resulting bioconjugates have thepotential to be used in a variety of applications such as imaging,diagnosis, and as combined photothermal and gene therapy forcancer treatment. Interestingly, the authors demonstrated thatadenovirus displaying an AuNP-labeling ratio of ∼2000:1 hadlimited effect on the infectivity of the adenovirus toward HeLacells; something that would not be expected a priori. As with manyrelated techniques, optimization of the conditions is critical foreffective separation.75



Analytical ultracentrifugation (AUC) utilizes a special type ofultracentrifuge that consists of a high-speed centrifuge rotor withcell compartments and an optical system (usually UV) tomeasure concentration gradients of the sample during theseparation process.76�78 The two principle modes used inAUC are sedimentation velocity and sedimentation equilibrium.While the technique can be applied to the purification of NM-bioconjugates, it is more commonly used for characterization ofthese composites and is discussed in more detail below.Dialysis and Filtration. After chromatography, dialysis, filtra-

tion, and the centrifugation-assisted versions of these techniques

probably represent the most common methods applied to NM-bioconjugate purification. Dialysis and membrane filtrationmethods in particular are relatively cheap, commercially avail-able, and simple to use. Both use some form of permeable orsemipermeable membrane with a given size or molecular weightcutoff (MWCO) value to separate the biomolecule from theNM-bioconjugate. In the case of dialysis, the membrane contain-ing the sample to be purified is immersed in a large excess volumeof liquid and species of MW < membrane MWCO (typically thebiomolecule) flow in the direction of high to lowconcentration.37,79 Filtration devices can be gravity driven, but

Figure 5. Chromatographic techniques and gel electrophoresis for characterization. (A) HPLC trace monitored at 210 nm of sample G (G5-Ac80-(NH2)109-alkyne2.7) shown with red dots. Five different peaks (0�4) were observed in the sample’s trace. Peak 0 had the same retention time asthe parent dendrimer (G5-Ac80-(NH2)32). Data were deconvoluted using peak fitting, individually fitted peaks are plotted in green, and the summationof the fitted peaks is plotted in blue. The fitting peak was developed to have the same shape as the parent dendrimer. Reprinted from ref 45. Copyright2010 American Chemical Society. (B) Size exclusion nanofluidics: (i) schematic of the size separation and metrology of a mixture of NPs by 3Dnanofluidic size exclusion, (ii) schematic of adjacent nanofluidic steps with excluded depths ds < dd ‘‘binning’’ NPs of different sizes ds < D < dd into asize subset. Schematics are not to scale. (iii) Etched channel surface as measured by scanning probe profilometry. Reprinted with permission from ref 46.Copyright 2010 The Royal Society of Chemistry. (C) Gel electrophoresis-characterization: (i) agarose gel characterization of maltose binding protein(MBP)-QD bioconjugates. Gel image clearly demonstrates the separation of QDs conjugates with different numbers of surface-assembled MBP. Reprintedfrom ref 64. Copyright 2006 American Chemical Society. (ii) Electrophoretic analysis of 5 nm AuNPs conjugated to polyT DNA 30�90 bases in length.Black arrowheads indicate visible bands as a guide to the eye. Reprinted from ref 42. Copyright 2008 American Chemical Society. (D) Ferguson analysis:(i) agarose gel picture showing BSP-coated 2.4, 4.8, 7.6, and 9.9 nm radius AuNPs (lanes A�D) visualized by absorption, and 540 nmQDs capped withDHLA (E), a 1:1 mixture of DHLA/DHLA-PEG600 (F), DHLA-PEG600 (G), and QD-20MBP-His5 conjugates (H). QD bands were visualized byfluorescence. (ii) Electrophoretic mobility of gold particle standards as a function of gel concentration. For each size, the corresponding retardationcoefficient, KR, is the slope of the linear fit. The inset shows KR1/2 as a function of the particle radius for the gold particles. (iii) Electrophoretic mobilityof QD samples vs gel concentration. Reprinted from ref 64. Copyright 2006 American Chemical Society.




generally centrifugation is used to increase the speed andefficiency of the process.80 For example, Vannoy assembledhen egg white lysozyme (HEWL)-QD conjugates using carbo-diimide (EDC) chemistry and then utilized 100 kDa MWCOAmicon ultrafilter devices to selectively remove the unboundHEWL (MW ∼ 15 kDa) along with excess reagents.80 Thepropensity of the resulting conjugates to form fibrils suggestedthat they could be used as a tool to understand fibril self-assemblyprocesses and the diseases associated with them. To date,ultrafiltration devices have shown some utility in purification ofbare NMs but also have good potential for application to NM-bioconjugates.81,82

Extraction. Chemical extraction and differential precipitationare other potentially efficient but largely underused purificationtechniques. In a relatively simple strategy that highlights thepossibilities available, Zhang and co-workers recently used abiphasic cap exchange-extraction method to generate amino-acidfunctionalized QDs, see Figure 4C.83 Water-soluble dithiocarba-mates, created through the reaction of carbon disulfide withprimary and secondary amine containing amino acids species,were dissolved in the aqueous layer of a biphasic reaction whileQDs with native hydrophobic capping ligands were present inthe organic chloroform layer. After overnight mixing of thereaction, the QDs had completely transitioned into the aqueousphase, demonstrating efficient cap exchange and extraction. Asonly QDs modified with surface biomolecules would transitioninto the aqueous phase, this extraction strategy could potentiallybe extended to purifying other QD or NP conjugates formedwith larger peptides or appropriately modified biomolecules suchas DNA, depending on their stability and tolerance to thereaction conditions. As an example of precipitation being usedfor purification purposes, B€ucking and co-workers used selectiveprecipitation of NP-bioconjugates by ammonium sulfate topurify BSA-coated InP/ZnS QDs.58 After the supernatant con-taining unbound protein and reagents was removed from theprecipitated NP-bioconjugates, they were redispersed in PBSbuffer prior to characterization. Both extraction and precipitationmethods typically rely on somewhat harsh conditions/solventsthat can be detrimental to biomolecular stability and are there-fore less commonly applied at the present time.

’CHARACTERIZATION

The ultimate test of successful NM-bioconjugation is of coursefunctionality in the desired application, and while the very natureof activity infers the presence and activity of the biomolecule onthe NM surface, this does not reflect any specific details of theunderlying NM-bioconjugate architecture. The latter may beessential to understanding any issues arising in subsequentbehavior. A variety of well-developed techniques which alreadyhave been successfully applied to characterization of bare NMsare now being extended to examining NM-bioconjugates. Theseinclude separation-based, scattering, microscopy, spectroscopy,mass spectroscopy, and thermal techniques. A discussion of thesetechniques along with tables summarizing select applications andsome relevant generalized details of each technique is detailedbelow.4,30�34

Separation Techniques. In general, the separation techni-ques summarized in Table 1 are relatively cheap and widelyavailable. As described above, some are already routinely used topurify NP-bioconjugates, and can also be quite effectively usedfor characterization purposes. Indeed they have been shown to be

particularly useful for confirming biomolecular attachment to theNM surface and providing information on approximate hydro-dynamic radius, NM-to-biomolecule conjugation ratio and in-sight into post-production degradation (i.e., stability).High-Performance Liquid Chromatography. As discussed

above, chromatography techniques, and in particular HPLCcoupled with anion exchange or reverse-phase columns, havedemonstrated the exquisite ability to resolve NM-bioconjugateswith different NM-to-biomolecule ratios, providing both thedistribution and overall average ratio of NM-to-biomoleculeper sample, see Figure 3A.22,42 In an especially challengingcharacterization example, Mullen and co-workers recently de-monstrated the use of reverse-phase HPLC to investigate poly-(amidoamine) (PAMAM) dendrimers modified with alkyneligands which are important for subsequent conjugation usingclick chemistry.45 Distinct peaks within each HPLC trace wereobserved, and deconvolution revealed the different dendrimer-ligand species allowing them to be quantified, see Figure 5A.From this analysis, the mean and distribution of dendrimer-ligand species within the sample were also determined, providinginsight into the actual sample composition.Nanofluidic Size Exclusion. Microfluidic and increasingly

nanofluidic devices have much to offer researchers from im-proved synthesis to characterization as well as reduced samplevolumes. Stravis and co-workers recently demonstrated a nano-fluidic size exclusion device for on-chip NP size separation andcharacterization, see Figure 5B.46 Currently the device is capableof separating particles in the 80�620 nm range with a step sizeresolution of 18 nm, although potential exists down to <10 nm.The authors used the device to characterize 100 and 210 nmfluorescent polystyrene particles and were able to determine sizedistribution profiles for each particle size. In the current context,this device has strong potential for purifying NM-bioconjugatesfrom both biomolecules and unmodified NMs while indepen-dently characterizing the size and providing information on theNM-bioconjugate distribution.Field Flow Fractionation. As discussed, FFF provides infor-

mation on charge, size (peak height position), and size distribu-tion (peak width).48,49 FFF has already been used by a number ofresearchers to provide size information on various unmodifiedNPs, including AuNPs,50 polymer (polyorganosiloxane-based)NPs,84 QDs/AuNPs,85 bionanocomposites of DNA/chitosan,51

and AuNPs modified with PEG ligands.50 As with many of thetechniques discussed, precalibration with known “size” NMstandards, such as Au or polystyrene NMs, is often desired. Sizestandard calibration is not always necessary, especially if thetechnique is combined with dynamic light scattering fordetection,50,86 or in the case of flow and sedimentation FFF, ifall geometric dimensions of the fractionation channel are accu-rately known.87,88 The latter allows you to convert retentiontimes into size distribution profiles (see refs 87 and 88 fortheory).Electrophoresis. A number of electrophoretic methods have

demonstrated wide utility in the characterization of NM-bioconjugates.52,53 For example, Gagnon and co-workers re-cently used dielectrophoresis to determine the effect that thesurface coverage of single stranded DNA (ssDNA) immobilizedonto silica NPs had on the ssDNA conformation and subsequentDNA hybridization efficiency.70 Gel electrophoresis in particularis routinely used to confirm bioconjugation by visualizing thediffering mobilities of NM-bioconjugates versus NMs.54�58,64,89

Under optimal conditions, gel electrophoresis has shown



exquisite resolution, being able to separate NMs biolabeled withone, two, or three biomolecules. This provides further informa-tion on the distribution of bioconjugation ratios obtained with aparticular chemistry, see Figure 5C.55,64 The technique can alsobeen used to elucidate conformational information about DNAimmobilized onto AuNPs.54

Ferguson analysis of gel electrophoretic data can be furtherused to infer the hydrodynamic diameter and ζ potential of NM-bioconjugates, as demonstrated for AuNP-DNA and QD-MBPbioconjugates.54,55,64 The analysis typically requires the user torun the sample in a series of agarose gels comprising a range ofconcentrations (i.e., 0.5�3.5%) and incorporate calibrationagainst known size standards prior to final analysis (such asAuNPs), see Figure 5D. The mobility of the unknown sample inthe gel series can be converted either to its retardation coefficientto yield the effective particle diameter or its ζ potential (see refs64 and 55 for some discussion of the theory along with goodexamples of the application). When Ferguson analysis is used to

evaluate gel electrophoretic data for the hydrodynamic diameterand ζ potential, a spherical assumption of the NM shape istypically assumed; however, with appropriate modifications it isalso suitable for the analysis of rod shaped NMs.90

Analytical Ultracentrifugation. In addition to purifying NM-bioconjugates, AUC can also determine the size, size distribution,and shape of NMs as well as structural and conformationalinformation about the conjugated biomolecules, see Figure 6A.AUC has been used for analyzing protein-based NMs such asthose assembled from cross-linked human serum albumin (HSA)along with inorganic NMs such as silica NPs and QDs.77,91�94

Falabella and co-workers recently used AUC to systematicallycharacterize single stranded DNA binding to 10 and 20 nmAuNPs as a function of surface loading and strand length, seeFigure 6A.95 Mulvaney’s group elegantly demonstrated thepotential power of this technique by characterizing QD sizedistributions, ligand densities and bioconjugation with the modelprotein BSA, all using AUC, see Figure 6A.93 They found the

Figure 6. Analytical ultracentrifugation and dynamic light scattering for characterization. (A) (i) Ultracentrifugation causes the NPs to sediment andultimately pellet in the bottom of the cell (nonuniform cell on right) relative to the reference. (ii) The instrument scans the absorbance from top-to-bottom of the sample and reference cells and absorbance is plotted versus cell position. The multiple curves represent absorbance scans at different timeintervals during the AUC process. Reprinted from ref 76. Copyright 2008 American Chemical Society. (iii) Progression of the sedimentation coefficientdistribution for bare 10 nm gold particles and 10 nm AuNPs modified with thymidine homo-oligomers decreasing from 30 bases (SH-T30) to 5 bases(SH-T5) in size. Reprinted from ref 95. Copyright 2010 American Chemical Society. (iv) Sedimentation coefficient distributions of 3.05 nm radiusCdSe/ZnS particles capped with DHLA-PEG in the presence (green circles) and absence (red circles) of BSA. Solid lines represent the fits of the data.The inset shows the change in average sedimentation coefficient with increasing numbers of bound BSA for different frictional ratio values. Reprintedfrom ref 93. Copyright 2008 American Chemical Society. (B) Dynamic light scattering (DLS). (i) The DLS and corresponding TEM characterization ofQD-luc8 NP-bioconjugates postpolymeric encapsulation. Polymeric encapsulation cross-linked two to three QD-luc8-NPs resulting in an increasedhydrodynamic radius observed in DLS. Reprinted with permission from ref 102. Copyright 2008 Elsevier. (ii) Hydrodynamic radii of QD-MBP-His5bioconjugates as a function of the average number of proteins per QD, measured from eight different samples. Schematics represent a CdSe-ZnS core-shell QD conjugated to ratios of 1, 2, and saturated with proteins. Reprinted from ref 64. Copyright 2006 American Chemical Society.




Table2.

Scattering

Techn

iquesa

technique/types

NM-bioconjugates

analyzed

advantages

disadvantages

refs

dynamiclight

scatterin

g(D

LS)

(alsoknow

nas

photon

correlation

spectroscopy

(PCS)

orquasielastic

light

scatterin

g(Q

RLS

))

QD-Luciferase

(8),QD-PEG

,

QD-M

BP

1.nondestructive.

2.hydrodynam

icdimensionsreadily

determ

ined.

3.rapid,simpleandrelativelycheap.

4.measurementscanbe

performed

inanyliquidmedia,

solventof

interest.

5.sensitive.

1.typically

assumes

sphericalshape.

2.resolutio

nissues.

3.biased

towardlargerparticles.

4.samplepreparationisextrem

elyimportant,

solutio

nsshouldbe

filteredpriorto

use.

5.lim

itedabilityto

measurepolydisperse

samples.

64,97�

103

fluorescence

correlation

spectroscopy

(FCS)

silverNP-DNA

1.hydrodynam

icdimensionsreadily

determ

ined.

2.photophysicalpropertiescanbe

determ

ined.

3.used

toinvestigatebindinginteractions.

1.lim

itedto

fluorescentspecies.

34,104�1

11

ζpotential

silicaNP-streptavidin,A

uNP-cytochrome

c,QD-[vario

uscappingagents]

1.measuresparticlestabilityandsurfacecharge.

30,64,113,114

Ram

anspectroscopy

resonanceRam

an(RR),Surfaceenhanced

Ram

an(SERS)

andSurface

enhanced

resonanceRam

an(SER

RS)

goldNP-hemoprotein,silverNP-

hemoprotein

1.provides

complem

entary

datato

IR.

2.isotopecompositio

ncanbe

determ

ined.

3.biom

olecularinteractions

canbe

investigated.

4.characteristic

fingerprintspectrum

obtained.

1.Ram

ansignalisrelativelyweakcompared

toRayleighscatterin

g.

2.lim

itedapplicationto

date.

115�

120

X-ray

diffraction(X

RD)

magnesium

alum

inum

layereddouble

hydroxidenanomaterial-D

NA,

lipidNP-vitaminA

1.polymorphicinform

ationdeterm

ined.

2.sizeandstructureofNM-bioconjugatecanbe

determ

ined,

dependingon

theNM

system

.

3.stabilityinform

ationcanbe

determ

ined

aboutthe

NM-bioconjugate.

1.lim

itedinform

ationobtained

aboutthe

nature

oftheNP-bioconjugateinteraction.

121�

128

smallangleX-ray

scatterin

g(SAXS)

AuN

P-DNA,polym

erNP-DNA,Q

D-

proteins,lipidNP-DNA

1.provides

macromolecularsize,shape,distances,

andhencepackingstructures.

2.canprovideaggregationinform

ation.

1.requirescrystalline

samples.

2.distinguishing

thesm

allanglescatterin

g

canbe

complex

(<10�).

3.biologicalentitiescanbe

damaged

bytheX-rays.

4.samplepreparationcanbe

complex

(thinfilmsrequired).

5.dataanalysiscomplex

andrequiresmodeling

tointerpret.

129�

134

smallangleneutronscatterin

g(SANS)

BSA

-polym

erNPs,PEG

-peptid

eNPs,

acrylamide-doxorubicinNMs

1.candeterm

inesize,shape

andorientationof

samples.

2.canbe

used

toprobe“soft”biologicalcontaining

samples

with

outd

amage.

3.isotopicsensitivitycanbe

used

toelucidatestructures.

1.lim

itedavailabilityof

neutronsources.

2.crystalline

samples

arerequired.

128,135,136

aGeneralinform

ation:

NM

structure,morphology,hydrodynam

icsize,aggregatio

nstate,biom

oleculeconformation,andNM-bioconjugatestabilitycanallbedeterm

ined.



sedimentation rate to be highly sensitive to particle size(resolution <0.16 nm), composition, and surface chemistry andwere able to model each individual contribution (assuming aspherical particle). Addition of BSA to the surface of the QDsresulted in a decrease in the measured sedimentation rate, withcurve fitting suggesting one-to-two BSA per QD in the finalbioconjugate. Benefits to the AUC method include small samplesizes, direct analysis of liquid-based dispersions without the needfor special solvents such as those commonly found in gradientcentrifugation, and amenability to a wide range of usable con-centrations. Often times, AUC is used in parallel with othercharacterization techniques, such as DLS and TEM;93,95 how-ever, as AUC is nondestructive to the sample it is possible toperform it and then sequentially apply a small portion to a moredestructive technique such as TEM.92

Electrospray Differential Mobility Analysis (ES-DMA). ES-DMA has previously been used to study a range of biomoleculesand unmodifiedNMs butmore recently has been extended to thecharacterization of NM-bioconjugates.96 Electrospray aeroso-lizes the sample which then passes through a neutralizer (to setcharge on the particles) into the DMA that separates positivelycharged particles based on their size-to-charge ratio where theyare detected using a condensation particle counter. ES-DMAthus provides the user an overall size and size distributionmeasurement based on the trajectory and the number of particlesand hence a concentration based on size. The technique wasrecently demonstrated by Pease and co-workers in conjunctionwith TEM for characterizing SA coated QDs and QD-functio-nalized with Lambda phage.96

Scattering Techniques.These techniques exploit the scatter-ing of radiation (e.g., light or energetic particles) through itsinteraction with a sample, see Table 2. Depending on thescattering technique applied, information about the NM struc-ture, morphology, hydrodynamic size, and aggregation state aswell as the biomolecular conformation and the NM-bioconjugatestability postproduction can be obtained.Dynamic Light Scattering. Of all the scattering-techniques

available, dynamic light scattering (DLS), also known as photoncorrelation spectroscopy (PCS), is probably the most commonlyused for characterizing the hydrodynamic size of NM-bioconju-gates, given that it is simple, noninvasive, nondestructive, andrelatively cheap to apply (the instrument itself can be costlythough).97�99 Fluctuations in the scattered light intensity, due tothe Brownian motion of the particles, are used to determine theparticle diffusion coefficient which is then related to its hydro-dynamic radius via the Stokes�Einstein relationship (see ref 64for an example of application).64,100 DLS has a broad workingconcentration and size range of ∼108�1012 particles/mL and10�1 000 nm, respectively; however, its sixth power dependenceon scattering versus particle size means that wide NM-bioconju-gate size distributions can obscure the presence of smallermaterials in the sample.100 In application, Murdock used DLSto study the effect of physiologically relevant dosing media (cellculture media with or without serum) on a variety of unmodifiedNMs (such as copper, silver, silicon dioxide, carbonnanotubes).99 This demonstrated that while many of the NMstended toward agglomeration in water, in some cases this effectwas reduced in serum containing cell culture presumably becausethe proteins present interact with the NM surface and mitigateagglomeration. Examples of direct NM-bioconjugate character-ization include examination of gelatin and human IgG adsorp-tion to AuNPs,101 luciferase (Luc8)-conjugated QDs,102 or

cytochrome P450 interactions with QDs,103 see Figure 6B. Jansand co-workers used DLS to characterize the nonspecific bindingof protein A to various size AuNPs and subsequently used it toquantitatively measure the binding of human IgG to protein Ain situ on the NPs at different incubation times determining thatat least two antibodies bound per protein A molecule on theAuNPs.98 Pons and co-workers used DLS to characterize thehydrodynamic radius and stability of CdSe/ZnS QDs modifiedwith various hydrophilic surface capping agents, including dihy-drolipoic acid (DHLA) and amphiphilic polymers along withself-assembled QD-MBP bioconjugates.64 They found that themeasured hydrodynamic size was strongly dependent on the coreQD size and the nature of the capping agent, with the DHLAligand producing much smaller-sized hydrophilic QDs than theirpolymer-coated or lipid encapsulated counterparts. The biden-tate DHLA were also found to produce more stable QDs thatwere less prone to aggregation compared to monothiol termi-nated ligands. Bioconjugation of QDs with MBP was monitoredvia DLS with the hydrodynamic radius found to saturate at ratiosexceeding 10MBP-per-QD, see Figure 6B. DLS is known to havepoor peak resolution and can only resolve particle populations(within the same sample) if they differ in size by at least a factor of3; therefore, DLSwould not be able to simultaneously resolve thedifferent MBP-per-QD ratios illustrated in Figure 6B if theycomprised a single sample.100 That said, the same data shows thatwhen analyzing separate purified samples (individually as-sembled and monodisperse) that clear differences in stoichiom-etry could be resolved between the different samples.64

Fluorescence Correlation Spectroscopy. Fluorescence corre-lation spectroscopy (FCS) is similar to DLS in that it measuressignal fluctuations due to diffusion, aggregation, interactions, etcand has already been successfully applied to accurately size thehydrodynamic radius of fluorescent NMs including QDs andfluorescent beads.34,104,105 With the use of confocal microscopy,a laser (in single or multiphoton excitation mode) interrogates asample containing a small number of particles and the fluorescentfluctuations observed within a confined optical volume are fittedto an autocorrelation function that can then be used to determinediffusion coefficients, see Figure 7A. The diffusion coefficient isagain correlated to the particle hydrodynamic radius via theStokes�Einstein equation. M€uller and co-workers investigatedthe use of dual-focus FCS to accurately size dye-doped latexparticles without precalibration to account for instrument re-sponse functions, the latter are usually dependent on theparticular instrument arrangement.105 The dual-focus systemused two laser beams to create two overlapping foci with a smallmeasurable shift between the two foci in the sample. Theresulting autocorrelation functions (ACF) for each foci and crosscorrelation function (CCF) between foci were used to determinethe NP radius. The authors found excellent agreement betweenthis and DLS for measuring different size particles. Crosscorrelation of data can also be performed with a single laser ifthe emitted light is measured in two different detection channels,as described by Roy and co-workers.106 In terms of NM-bioconjugates, FCS has also been used to determine bindingkinetics between 100 nm unilamellar vesicles and fluorescentlylabeled peptides.107 The correlation times for the free peptideversus the peptide bound to the large unilamellar vesiclesinvestigated in this format differed significantly, which allowedfor accurate kinetic measurement even at low nanomolar peptideconcentrations, see Figure 7A. This technique has even beenextended to observing metal enhanced fluorescence (MEF)



Figure 7. Continued




resulting from Cy5-labeled DNA hybridizing to DNA-modifiedsilver NPs by using the modified technique of fluorescencelifetime correlation spectroscopy.108 This derivative approachmeasures intensity decay times, instead of intensity fluctuations,and found a 5-fold decrease in emission lifetime of the Cy5-DNAupon binding to its complement attached to 50 nm silver NPs.Hybridization also resulted in a 10-fold increase in Cy5 fluores-cence intensity and increased its contribution to the autocorrela-tion function compared to free Cy5-DNA. These effects allowedthe researchers to resolve different species within the samplemixture based on a combination of both intensities and lifetimes.A similar approach was used by Tang and co-workers to measurealpha fetal protein (AFP).109 Here, silver NPs were functiona-lized with anti-AFP and AlexaFluor647-labeled AFP was used asthe tracer antigen in a homogeneous competitive assay for AFP.Popular derivatives of FCS include dual-color FCS, which

allows the user to cross-correlate data from two differentfluorescent channels simultaneously110 and F€orster resonanceenergy transfer (FRET)-FCS, commonly called single-pair orparticle (sp)FRET. Pons and co-workers used single-particleFRET measurements taken on a dual color FCS system toinvestigate the binding of acceptor-dye-labeled MBP to QDs(donor).111 Donor and acceptor signals were separated using adichroic filter, and results demonstrated that protein binding toQDs followed an expected Poisson distribution. Since FCS is,for all intents and purposes, a single molecule technique, animportant factor often overlooked in performing this techniqueis getting the final sample concentration into the correct

dilution regime, with typical concentrations in the low tosubnanomolar range.Resonance Light Scattering Correlation Spectroscopy.

RLSCS measures the resonant light scattering intensity fluctua-tions from NMs that posses a surface plasmon band (such as Auor silver NPs). Although limited to suitable NP materials, thetechnique has been used to study the bioconjugation of AuNPswith either BSA or thiol-modified DNA species, with the inter-action resulting in increased diffusion times.112

Zeta (ζ) Potential.Measurement of the ζ potential of a NM insolution provides information on the net charge a NM-biocon-jugate has and provides insight into NM-bioconjugate stability.The ζ potential is commonly determined by applying an electricfield across a sample andmeasuring the velocity at which chargedspecies move toward the electrode; this is proportional to the ζpotential.64 The ζ potential can also be used to infer particlestability with a value of (30 mV often selected as an arbitrarydelineation of stability. Values >30 mV indicate stability, whilevalues <30 mV represent particles with a tendency towardagglomeration or instability.30 Many factors can influence NMstability (and hence the ζ potential) including pH, concentra-tion, ionic strength of the solution, temperature, radiation, andthe nature of the surface ligands. ζ potential measurements havebeen used to study the stability and particle size of SA-functio-nalized silica NPs as a function of pH,113 the surface coverage ofcytochrome c bioconjugated to AuNPs,114 and various QDcapping agents.64 In the latter AuNP example, Gomes and co-workers used ζ potential measurements to evaluate the surface

Figure 7. Optical and electron microscopy techniques: (A) (i) state-of-the-art fluorescence detection in solution based on a confocal microscope andfluorescence correlation analysis. The inset shows the size of the typical analysis volume. Reprinted with permission from ref 141. Copyright 2010MDPI.(ii) Peptide binding vesicles. Time-resolved count rates from solutions containing 2 nM Alexa-labeled peptide only (top trace), Alexa-labeled peptideplus 0.2 μM lipid vesicle (middle trace), and Alexa-labeled peptide plus 1 μM lipid vesicle (bottom trace). (iii) The effect of vesicle (lipid) concentrationon the autocorrelation curves obtained from a solution containing 2 nM peptide. Autocorrelation curves are shown from 0 to 4 μM accessible lipidconcentrations that correspond to 0�90% peptide bound. Reprinted with permission from ref 107. Copyright 2004 Elsevier. (B) Electron microscopy:(i) low-resolution TEM images of gold conjugates. Image shows a thin white “halo” layer surrounding the surface of the NPs indicating coating of thegold with protein. This “halo” effect is not present on the nonconjugated AuNP image, see insert. Reprinted with permission from ref 150. Copyright2010 Elsevier. (ii) Uranyl formate staining of DNAOrigami. Orthographic projectionmodels and TEMdata of two icosahedron particles. The scale barson TEM images are 100 nm. Reprinted with permission from ref 152. Copyright 2009Macmillan Publishers Ltd. (iii) Uranyl acetate staining of CPMV-C60 fullerene conjugates for TEM imaging. Reprinted from ref 151. Copyright 2009 American Chemical Society. (C) Atomic forcemicroscopy of carbonNTswrapped with (GT)30 oligonucleotide. (i) Height image (5 nm scale) and (ii) phase image (25� scale) of one representative carbonNT-DNA, alongwith (iii) a 3D representation (1.7 nm scale) indicating the peak height (p), valley height (v), peak width (w), and pitch (pi) measurements. (iv)Distributions of nanotube height measurements at peaks and valleys (n = 300 CNTs). Reprinted from ref 164. Copyright 2008 American ChemicalSociety. (D) AFM topographic data obtained for the measurement of (i) nonconjugated DHLA-QDs and (ii) EYFP-conjugated DHLA-QDs (ratio 10).The resulting histogram from particle cross sectionmeasurements is shown on the right. The inserts represent the corresponding cross section profiles ofthe dotted line in the AFM-height image. Reprinted with permission from ref 165. Copyright 2009 Wiley-VCH Verlag GmbH & Co. KGaA. (E) DNAOrigami and AFM analysis. Folding circular genomic DNA from the virus M13mp18 into different shapes. Models are shown in the top row, with theresulting AFM images shown below. Reprinted with permission from ref 168. Copyright 2006 Macmillan Publishers Ltd.




Table3.

Microscop

yTechn

iquesa

technique/types

NM-bioconjugates

analyzed

advantages

disadvantages

refs

atom

icforcemicroscopy(AFM

)AuN

P-DNA,carbonnanotubes

(CNT)-DNA,Q

D-streptavidin,

titanium

dioxide(T

iO2)

NP-DNA,

AuN

P-cytochromec

1.capableof

3Dmapping

ofthesamplesurface.

2.cananalyzeindividualNMs.

3.canbe

appliedto

nonconductingwetandsoft

samples

onvario

ussubstrates.

4.size

andshapecharacterizationof

theNM

and

biom

oleculecanbe

determ

ined.

5.biom

olecularinteractions

canbe

characterized

usingfunctio

nalized

tips.

1.sampleunderstudymustbe

immobilized

onto

asuitablesurfaceandsamplepreparation

canbe

complex.

2.onlysm

allareascanbe

mappedandthe

scan

timecanbe

slow

.

3.analysisis,ingeneral,lim

itedto

theNPexterio

r.

4.techniquerequiresextensiveoptim

ization

andinterpretatio

n.

114,149,164�

169

transm

ission

electron

microscopy

(TEM

)

cryogenic-TEM

allowsanalysisof

tissuesamples

AuN

P-hemoproteins,polylactideNP-

adenoviru

ses,lipidmicelles,vesicles

andbiolayers-transient

nanostructures

1.cananalyzeindividualNMs,canbe

used

to

determ

ineNM

size

andshapes.

2.aspectratio

scanbe

determ

ined.

3.biom

olecules

canbe

somew

hatvisualized.

1.conductin

gor

stainedultrathinsamplerequired.

2.drysamples

needed

foranalysis,nonphysiological.

3.ionizing

radiationcancausesampledamage.

4.sm

allangleof

view

,onlylim

itedNPs

can

beanalyzed

atonetim

e.

5.expensiveequipm

entand

technical

expertiserequiredto

obtainmeaningfuldata.

61,148�1

53

scanning

electron

microscopy

(SEM

)

environm

entalSEM

(ESE

M)

polystyreneNP-protein

1.cananalyzeindividualNM

core,can

beused

to

determ

ineNM

size

andshapes.

2.candeterm

ineNM

compositio

n.

3.biom

olecules

canbe

imaged

usingES

EM.

1.conductin

gsampleusually

required.

2.drysamples

needed

foranalysis,

nonphysiological(unlessrunn

ingES

EM).

3.expensiveequipm

entand

technical

expertiserequiredto

obtainmeaningfuldata.

4.mainlyused

tocharacterizeNPcore.

5.ES

EMreducesresolutio

n.

154�

160

light

microscopy

standard

light

microscopyand

fluorescence

confocalimaging

silverNP-DNA,lum

inescent

NP-proteins

1.single-m

oleculebasedmeasurementsarepossible

with

fluorescentsamples.

2.morewidespreadthan

manyof

theother

microscopybasedtechniques.

1.somew

hatlimitedby

diffractionissues.

2.canrequire

fluorescentsamples.

139�

146

near-fieldscanning

optical

microscopy(N

SOM)

Au-NP-streptavidin-biotin

1.high-resolutionsurfaceanalysisatam

bientconditions.

2.contrasttechniques

availablewith

optical

microscopymay

also

beappliedforhigh

resolutio

n

analysis.

1.sm

allsurface

area

analyzed.

2.analysislim

itedto

thesurfaceof

theNP.

3.currently

a“specialist”technique.

170,171

aIn

manycases,capableof

singleparticleresolutio

n.Techniquescanbe

used

todeterm

ineNM

size

andshape.



coverage of heart- and yeast-cytochrome c on 11.5 nm goldcolloids following overnight incubation.114 They determined thatthe change in ζ potential depended on the [cytochromec]/[AuNP] ratio in the sample mixture, and this was found tobe saturated for both proteins at ratios of ∼200:1 producingstable NP-bioconjugates with a final ζ potential ranging from�30 to �35 mV. As with DLS, the ζ potential is fairly easy tomeasure and oftentimes the two techniques are available on thesame instrument.Raman Techniques. Raman techniques commonly applied to

characterization include Raman spectroscopy and surface en-hanced Raman scattering (SERS). Raman spectroscopy mea-sures the inelastic scattering of monochromatic radiation (UV,visible, or near IR) by a sample. The incident light becomes eitherStokes or anti-Stokes shifted in wavelength resulting in sharpfingerprint Raman bands that are characteristic of the sample andcomplementary to infrared (IR) spectroscopy. Carbon nano-tubes, for example, exhibit strong Raman scattering which hasbeen found to be sensitive to isotope composition115 andbiomolecular interactions.116 Yang and co-workers used Ramanspectroscopy to investigate the noncovalent interaction betweenpyrene-labeled cellulose (Py-HPC) and multiwalled carbonnanotubes (MWCNTs).116 Pristine MWCNTs had character-istic Raman D (C�C) and G (CdC) bands at 1321 and1565 cm�1, respectively. Enhancement of the D band wasobserved upon binding Py-HPC, not only confirming bioconju-gation but also the nature of the interaction, since D bandenhancement is attributed to π�π stacking.Metallic NPs and in particular silver NPs offer the unique

possibility of localized surface enhanced Raman scattering(SERS)117,118 and have been used to study hemoprotein (heme-containing proteins such as hemoglobin, myoglobin, and cyto-chrome c) and lysozyme bioconjugation to Au and/or silver NPsalong with aptamer conformational changes upon targetbinding.119,120 Zhang and co-workers looked at the interactionof lysozyme upon noncovalent interaction with 90 nm AuNPs,which induced protein aggregation resulting in the formation ofextended protein-NP assemblies.120 A blue shift in the amide Iband combined with a red shift in the amide II band of the Ramanspectra revealed conformational modifications in the gold-boundlysozyme. This was attributed to a shift from theR-helical structureto a more β-sheet or random coil conformation.X-ray Diffraction. X-ray diffraction (XRD) is typically used to

provide structural information about crystalline samples. It is alsofrequently used to characterize materials containing nanosizedcomponents embedded in an extended biological matrix, such asthose found in tissue scaffolds and bone cements121,122 ornanobioconjugate layered materials such as nanobiohybrids,123,124

where analysis of d-spacing changes upon bioconjugation be-tween layers of the nanomaterial can be used to assess reactioncompletion or investigate the biomolecular orientation.125,126

For example, Choy and co-workers prepared anionic clayscomprising brucite-like cationic hydroxide nanolayers with ex-changeable anions.126 When the nitrate ions were replaced withDNA, the spacing between hydroxide layers was found toincrease ∼2 nm consistent with the DNA double helix layingparallel to the basal plane of the layers. In some cases, XRD isused to characterize the biomolecule after it has been associatedwith theNM, as seen in its use to assess the polymorph stability ofsolid-lipid NPs containing vitamin A.127 XRD had also been usedto assess the effect of PEG content on the self-assembly ofpeptide fibril nanostructures.128

Small-Angle Scattering Techniques. Small-angle X-ray scat-tering (SAXS) and small-angle neutron scattering (SANS) havebeen applied to elucidating information on the structure, mor-phology, and characteristic intra-assembly spacings of a variety ofpolymer and biological based nanomaterials.129,130 The principlemechanistic difference between SAXS and SANS is that X-raysare scattered by the electrons while neutrons are scattered by thenucleus. This has a direct impact on what can be observed and themagnitude of the observation (e.g., in neutron scattering the scat-tering factor of carbon and hydrogen is of approximately equalmagnitude but opposite sign while with X-rays there is about asix-fold difference between carbon and hydrogen). Both X-raysand neutrons can be used at a variety of wavelengths; importantlythough radiation damage to the sample is a function of wave-length (i.e., energy) and exposure time. For biological samples,the X-ray radiation used is generally 1.54 Å but can range fromabout 0.2 to 4.0 Å. It should be noted that long-wavelength X-rays are absorbed strongly by airand thus are seldom used.Neutrons used to study biological materials are generally in therangeof 2�10 Å (the full spectrum of neutron radiation is fromabout 1.0 to 20 Å).SAXS has been used for analyzing various NM-bioconjugate

systems including the packing and extended structures of DNA-modified AuNPs,131 elucidating the internal structure of heparinloaded chitosan NMs,132 and determining the structural changesthat occur in various ordered mesoporous (nanoporous) materi-als upon adsorption and release of lysozyme.133 Bhattacharyyacompared two different ordered mesoporous materials, silica-based SBA-15 and organosilica-based MSE, during their adsorp-tion and release of lysozyme.133 SAXS analysis revealed that themore hydrophilic material SBA-15 produced lattice structureincreases in size upon adsorption and also release of lysozyme,whereas the MSE lattice structure remained constant duringthese conditions resulting in a slightly different release profile. Arecent study by McKenzie and coworkers demonstrated amicroscale flow system for simultaneous in situ monitoring ofSAXS and UV�visible spectra.134 The system was used for thereal-time size determination of mercaptoethoxyethoxyethanol-stabilized AuNPs (0.8�5 nm) and while no bioconjugates wereinvestigated here the technique could quite readily be applied tothe characterization of NM-bioconjugates. Several studies haveused SANS in combination with DLS and static light scattering(SLS) to elucidate the structure and core morphology ofNMbioconjugates.128,135,136 Serefoglou and coworkers, for ex-ample, used SANS to study complexes formed between theprotein BSA and two anionic graft copolymers, revealing core-corona NPs whose core size increased with decreasing content ofthe neutral poly(N,N-dimethylacrylamide) side chains in thegraft copolymers.135 However, the SANS technique suffers froma lack of suitable neutron sources (synchrotrons) and is thus notwidely utilized at this time.Microscopy. These techniques are based on visualizing a

sample using light, electrons, or a scanning probe, as summarizedin Table 3.137�139 Traditional optical light microscopy (some-times referred to as bright field microscopy), and its myriad offluorescent derivatives were until recently, however, unable toresolve nanoscale features <200 nm due to diffraction limitations.As a result, they found limited application in the characterizationof individual NM-bioconjugates, although they have been ex-tensively used to track and image an almost countless number offluorescent NM-bioconjugates in cellular and small animalstudies. Recent developments involving ultrahigh-resolution



Table4.

SpectroscopicTechn

iquesa

technique/types

NM-bioconjugates

analyzed

advantages

disadvantages

refs

UV�v

isiblespectroscopy

hasbeen

used

tocharacterizealarge

numberof

NM-bioconjugatetypes.

Exam

ples:A

uNP-protein,AuN

P-DNA,ironoxideNP-protein,

QD-protein

1.costeffective,availableinmostlaboratories,

simpleandfast.

2.anumberof

NMshave

intrinsicoptical

propertiesthatcanbe

used

todeterm

ine

concentration,size

andsometimes

shape.

Changes

inthespectracanbe

associated

with

aggregationstate.

3.biom

olecules

also

typically

have

someintrinsic

opticalpropertiesthatcanbe

used

todeterm

ine

concentration.

4.canbe

used

tocharacterizeNM-biomolecule

conjugationandaverageratio

.

1.requiresfairlyconcentrated

samples

assensitivityislow.

2.provides

averageinform

ationonlyno

distrib

utioninform

ation.

4,114,119,

173�

184

circulardichroism

(CD)

UVto

visibleCD,vibratio

nalC

D,

andsynchrotronCD

AuN

P-protein,silicaNP-protein,iro

noxideNP-protein,CNTs-protein,

QD-protein

1.provides

biom

olecule(proteinandDNAmost

common)conformationaland

stabilityinform

ation.

2.nondestructive.

3.liquidsamples

canbe

prepared

inphysiological

environm

ent.

1.extensivecharacterizationof

the

biom

oleculepriorto

NPbioconjugatio

nrequiredto

obtainmeaningfulresults.

2.nonUV-absorbing

buffersrequired.

3.oxygen

canbe

interferentinUVanalysis.

4,59,94,

114,179,

182,185�

190

fluorescence

spectroscopy

standard

fluorescence,w

avelength

andtim

e-resolved.B

othdirectand

indirecttechniques.fluorescence

resonanceenergy

transfer

(FRET

):specialfluorescentform

atthat

requiresadonorfluorophoreandan

acceptor

species.

AuN

P-BSA

,silica

NP-β-

lactoglobulin,A

uNP-DNA,Q

D-

dopamine,lanthanide

doped

siloxane

NP-R-bungarotoxin,

QD-M

BP

1.sensitive.

2.fluorescence

canbe

environm

entally

sensitive

which

canbe

used

toprovidebiom

olecule

conformationalinformationor

confirm

attachmentto

NM.

3.NM-biomoleculeaveragecouplingratio

,insome

instancesdistrib

utioninform

ation.

4.FR

ETform

atsensitive

tomoleculardistances.

1.intrinsicor

extrinsicfluorescence

required

which

may

necessitatelabeling.

37,89,145,

146,182,

186,187,

191�

209

infrared

spectroscopy

(IR)

standard

IF,Fouriertransform-IR

(FT-IR)andattenuated

totalreflectio

nIR

(ATR-IR)

diam

ondNP-biotin,A

uNP-

dextran/albumin,A

uNP-

hemoproteins,silver

NP-hemoproteins,

silicon

NP-streptavidin,

silicaNP-β-lactoglobulin

1.confirm

sNM-biomoleculeattachmentthrough

theappearance

ofcharacteristic

fingerprintIR

bands.

2.canprovideconformationalinformationaboutthe

biom

oleculeupon

NM

bioconjugatio

n.

1.samplepreparationcanbe

complicated.

2.optim

izationof

thetechniqueis

oftenrequired.

3.H2O

isan

interference,strongabsorbance.

4.appropriatebackgroundsarerequired.

67,119,

179,172,

180,182,

186,210�

213

nuclearmagnetic

resonance

(NMR)b

2DNMR,N

OENMR,T

ROSY

NMR

silicaNP-human

carbonicanhydrase,

lipidNP-PE

G,A

uNP-PE

G,

dendrim

erNP-surfactant

species

1.nondestructivetechnique.

2.physical,chemical,structural,andenvironm

ental

inform

ationavailableaboutthebiom

olecule.

1.onlycertainnucleiareNMRactive.

2.relativelyinsensitive,requireshigh

sample

concentrations.

3.deuterated

solventsrequired.

4.techniquerelativelyexpensiveand

requiresexpertise.

45,64,172,

211,214

aIncludesavarietyoftechniqu

esthatcanprovideNM-biomoleculeconjugationconfi

rmation,conformationalinformationaboutthe

immobilizedbiom

olecule,NM

size,and

insomeinstances,shape,

averageNM-biomoleculeratio

,stability,andconcentration.

b2D

NMR,2-dimension

alNMR;NOENMR,nu

clearOverhausereffectNMR;TROSY

NMR,transverse

relaxatio

nop

timized

spectroscopy.



optical systems especially in fluorescent microscopy now providethe ability to resolve features less than 100 nm.139�142 Inaddition, single-molecule microscopic measurements of fluor-escent species and especially QDs have become more com-monly used to study various biological processes.143 Othertypes of pertinent microscopic analyses that have been accom-plished include monitoring Cy5-labeled DNA hybridizing toDNA-modified silver NPs,144 stepwise photobleaching offluorescently labeled (Alexa488) proteins attached to lantha-nide-ion doped oxide luminescent NPs,145 or dye-labeledssDNA binding to silica NPs146 along with various F€orsterresonance energy transfer (FRET) studies.143 These are dis-cussed further in the fluorescence and emerging technologiessections below.Electron Microscopy. TEM and SEM readily obtain single

particle resolution and are more frequently applied to character-izing the NM core along with core-shell size and/or structure.Both techniques rely on the wave nature of electrons to directlyilluminate a sample in either transmission or reflectance modes,respectively, and generate an image.137 The direct imagingcapability of the TEM and SEM is particularly useful for NMswith nonspherical shapes allowing a direct measure of aspectratio.119,147 While TEM is more commonly used to image thecore NM itself, it is possible at low accelerating voltages tovisualize biomolecules attached to a NM core as demonstratedfor polylactide NM-adenoviruses constructs,148 carbon NTsmodified with horse radish peroxidase or antibodies,149 andAuNPs modified with antibodies,150 see Figure 7B. As Thobhanipointed out, this layer appears to be much smaller than expectedfrom concurrent DLS analysis of the bioconjugated-NMs and islikely a result of the drying required for TEM analysis.150 Imagingthe biological components can be facilitated by staining thebiomolecules with contrast agents such as uranyl acetate oruranyl formate as demonstrated for Potato virus X, fullerene-labeled CPMVNPs, and DNA-derived 3D nanoscale shapes (seeFigure 7B) but, as with drying, staining may also alter the sizes ofthe NM-bioconjugate.61,151,152 Although dried samples are typi-cally required for TEM analysis, researchers recently imagedliquid samples of fixed fibroblast cells stained with epidermalgrowth factor (EGF)-labeled AuNPs by scanning TEM using aspecially designed microfluidic device.153 Since SEM uses elec-trons to image a surface in reflectance mode, it is mainly used toimageNM corematerials154�156 but has the added advantage of alarger imaging field of view than the TEM.147,156 SEM has lessresolving power for features <20 nm than TEM, although thetechnology is steadily improving even in this area. SEM is oftencombined with energy-dispersive X-ray spectroscopy (EDX)analysis to confirm elemental composition of the NM.157,158

Environmental SEM (ESEM) does allow sample imaging underlow pressure, fairly high humidity, and without the requirementof a conducting overcoat but has found limited use to date for thestudy of NM-bioconjugates.159,160

Atomic Force Microscopy. In contrast to TEM and SEM,atomic force microscopy (AFM) is a scanning probe techniquethat can divulge a range of information about the NM, thebiomolecule, and the NM-biomolecule interaction on a single-particle basis. Although once considered a fairly specializedtechnique, recent advances in both instrumentation and softwarehave not only expanded the technique’s capabilities but havealso made it more user-friendly. Unlike TEM and SEM, whichare best performed with conducting or semiconductingsamples under vacuum conditions, AFM can be applied to

nonconductive, wet, and soft samples, allowing for many differ-ent types of materials to be analyzed in physiologicalenvironments.138,161,162 Samples can be analyzed using variousscanning modes including static (noncontact) and dynamic/tapping (contact and intermittent sample contact), which pro-vide information about various sample parameters includingmorphological information about the NP (size and shape, etc),information about the biomolecule (adhesion, elasticity,Young’s modulus, and stretching parameters), and interactionsbetween the individual particles and biomolecules (proteinfolding-unfolding, biomolecular attachment to a NM surface,NM-NM agglomeration, etc.).162,163 AFM has been applied tostudying the interactions between multiple NM platforms andbiomolecules including QDs, and carbon NTs binding toDNA,164,165 as well as AuNPs, carbon NTs, and QDs functio-nalized with different proteins,114,149,165,166 see Figure 7C. Thesensitivity and versatility of AFM has allowed researchers tomap ligands attached to a surface in 3D30 and interestingly,when this is combined with data clustering analysis, it canprovide approximate NM-bioconjugate stoichiometry, seeFigure 7D.165 Thus, AFM could potentially allow for a 3Drepresentation of ligands or biologicals attached to a NMsurface, the determination of how strongly bound the ligandsare to that surface, as well as the structure of the under-lying NM core. AFM has also been particularly useful incharacterizing DNA derived nanostructures, so-called DNAOrigami, which is based on programmed self-assembly ofspecially designed branched DNA junctions and tiles, seeFigure 7E.152,167�169

Near-Field Scanning Optical Microscopy. Another surfaceprobe microscopy technique, near-field scanning optical micro-scopy (NSOM also known as SNOM), breaks the opticalresolution limit of light microscopy by placing the detectionprobe extremely close (at distances much smaller than thewavelength of light) to the surface to be analyzed. High-frequency spatial and spectral information is obtained by analyz-ing the evanescent fields close to the sample surface. NSOM hasthe advantage of combining optical and/or spectroscopic datawith high-resolution surface topographical information. Contrastproperties, such as phase contrast, polarization, fluorescence,staining, etc., that are available through traditional opticalmicroscopy are also available with NSOM. These propertiesallow for a variety of chemical interactions (such as biomolecularbinding, ion sensing, nearfield surface enhanced Raman spec-troscopy, etc.) to be studied at high spatial resolutions. Suchversatility and specificity can also allow for a variety of NM-bioconjugate properties to be studied. For example, interactionsbetween AuNP-streptavidin (SA) bioconjugates and biotin-functionalized surfaces have been studied using this technique.170

NSOM of the surface was performed with a 514 nm argon laserattached to the end of an optical bent-fiber probe and scatteredlight from the AuNP detected via an avalanche photodiode.AuNP-SA bioconjugates were clearly observed in biotin-mod-ified regions of the surface confirming the specificity of the SA-biotin interaction. Other interactions between the biomoleculesand NMs can also be studied including the ability to differentiatebetween specific versus nonspecific interactions. NSOM/QD-based dipole-emission fluorescence imaging has been used tostudy the nanostructure of antigens on Yersinia pestis vaccineparticles comprising V immunogen fused with protein anchor(V-PA) loaded on a gram positive enhancer matrix (GEM).171

Here, Zeng and co-workers used a combination of biotin-labeled



anti-c-Myc (which recognizes the c-Myc tag expressed on theV-PA fusion protein) and SA-conjugated QDs for immunostain-ing of the V-PA loaded GEM vaccine particles. They found thatthe V-PA tended to bind to the GEM surface in the form ofnanoclusters, containing multiple (two or more) V-PA mol-ecules, rather than single V-PA monomers, with ∼3000 of theseV-PA immunogen structures bound to a single GEM particlesurface. Such high density of these immunogen structures isadvantageous as the resulting V-PA-GEM vaccine particles arelikely to elicit a robust immune response.Spectroscopic Techniques. These techniques exploit the

interaction of electromagnetic radiation with a sample materialresulting in the wavelength dependent absorption, and in the caseof fluorescence re-emission, of radiation, see also Table 4.Typically, a wavelength dependent spectrum is produced withcharacteristic absorption/emission peaks inherent to thesample.172 Spectroscopic techniques can provide average bulkanalysis of the NM-bioconjugate including information concern-ing the confirmation of successful NM-biomolecule conjugation,the conformational state of the biomolecule once attached to theNM, the average NM-to-biomolecule ratio, and the stability ofthe resulting NM-bioconjugate.UV�Visible Absorbance. The intrinsic UV�visible absor-

bance properties of many NMs can be used to monitor pertinentproperties, such as concentration, size, and aggregation state.QDs for example have size-sensitive absorption profiles,173 andoptical absorption has been used to characterize composition,size, and purity of single-walled carbon NTs.174 Metal NPs, inparticular Au and silver, exhibit a strong absorption in the visibleregion known as the surface plasmon resonance (SPR) band(often shortened to the SP band). The SP band is dependent on anumber of factors and is found to be sensitive to size, shape,composition (i.e., silver, Au, nanoshell structures), aggregation

state, and also refractive index changes within surfaceproximity.114,119,175,176 A number of researchers have looked atthe effect of proteins binding to gold and silver NPs, finding small(∼5 nm) shifts in the UV�visible measured SP band as a resultof protein binding and in some cases larger shifts due tosubsequent aggregation.114,119,177,178 These changes alone cansometimes serve to confirm NM interactions or bioconjugation.Both direct and in-direct analysis of UV�visible spectroscopy

data can also provide information on the NM-bioconjugate.Direct characterization is possible when the biomolecule has adistinct UV�visible profile that remains discernible upon con-jugation to the NM. Protein absorption bands at 280�290 nmand the soret bands (410 nm) of hemoproteins such as cyto-chrome c have been used to directly quantify the average amountof protein immobilized on a NM surface.4,179 Even if the NMdisplays a significant level of absorption, the presence of newpeaks or increases in certain regions can confirm the presence offor example DNA (Abs∼260 nm). Proteins coeluting with NMscan also be viewed in gel electrophoresis using colorimetricstains, such as Coosamine Blue, which can then be quantifiedwith appropriate instrumentation (such as a fiber optic spectro-photometer) to determine the average NM-to-biomoleculeratio. Alternatively, with the use of a more indirect method, theamount of protein present in solution before and after NMexposure can be quantified using either protein absorbance at280�290 nm180�182 or a number of reactive colorimetric assays,including the Bradford reagent and bicinchoninic acid (BCA)assays,183 as demonstrated for gold-coated magnetic particlesmodified with antibody fragments.184 Many of these tests,although not specifically designed for NMs, can be applied totheir analysis. However, as NMs have been demonstrated tointerfere with certain chemical assays and tests, appropriateanalysis of controls and the “naked” particles should be

Figure 8. Spectroscopy: (A) CD spectra of unlabeled (top) and AuNP-labeled cytochrome c proteins (middle and bottom), with no salt (middle), with0.1 M NaCl (bottom). Specific amino acids mutated to cysteine to allow subsequent NP labeling are indicated. Reprinted with permission from ref 59.Copyright 2009 PNAS. (B) (i) Infrared spectra of lysozyme attached to complex nanodiamonds (cNDs) of various sizes: (1) Lyso-5cND; (2) Lyso-50cND; (3) Lyso-100cND; (4) Lyso-200cND; (5) Lyso-300cND; (6) Lyso-400cND; (7) Lyso-500cND; (8) lysozyme; (9) 100cND. (ii) Amide Ifrequencies of the attached protein lysozyme as a function of cND size. Reprinted from ref 180. Copyright 2011 American Chemical Society. (C) FT-IRspectra of QDs with different surface ligands: (top) DHLA-PEG600; (middle) DHLA-PEG600/DHLA-PEG400-biotin (4:1 molar ratio); (bottom)DHLA-PEG600/DHLA-PEG400- COOH (19:1 molar ratio). (I) and (II) can be assigned as amide I and amide II bands, respectively. Reprinted fromref 213. Copyright 2007 American Chemical Society.




undertaken prior to interpretation of the results. While allowingderivation of an average NM-to-biomolecule ratio, such analysis,however, does not give a NM-bioconjugate product distributionprofile or information regarding biomolecular conformationonce attached to the NM surface.Circular Dichroism. Conventional circular dichroism (CD)

measures the ability of optically active materials to differentiallyabsorb circularly polarized light (usually UV) and is commonlyapplied to the conformational analysis of biomolecules such asproteins and nucleic acids.4,185 Far-UV (<250 nm) CD providesstructural information concerning the proteins' secondary struc-ture and has been successfully used to studying protein con-formational changes that occur following both initial adsorptionand later stable interactions with a variety of NM surfaces,including Au,59,114,179,186 silica,94,182 carbon NTs,187 iron oxide,188

and QDs.189 Importantly, the observed conformational changesof proteins interacting specifically or nonspecifically with NMsurfaces have been found to also depend on a number of otherfactors including pH,186 surface density of the protein,182 addi-tional ligands present on the NM surface, and temperature.4

Aubin-Tam used CD to monitor the global structure of cyto-chrome c as it was labeled with a 1.5 nm diameter AuNP atvarious mutational sites placed throughout its structure, seeFigure 8A.59 Losses in protein helicity were correlated with grosschanges in structure, and these were further confirmed bymodeling (see also below). A number of recent developmentsin CD technology offer some exciting prospects for extendingthis analysis.185 For example, synchrotron radiation CD (SRCD)offers more detailed spectral information at wavelengths<190 nm due to its high light flux in this region, although themain disadvantage is limited availability. Vibrational CD (VCD)uses polarized light in the infrared (IR) region, looking pre-dominately at variations in the amide I and amide II bands ofproteins, although it does require replacement of H2O with D2Oin solution-based measurements due to the strong absorption ofwater in this region.190

Fluorescence Spectroscopy. This technique in both steady-state and time-resolved modes offers a powerful and sensitivetechnique for determining a number of the parameters associatedwith the immobilization of biomolecules to a NM surfaceincluding fluorophore local environment, biomolecule-NM cou-pling ratio, conformational state, and in some instances, intra-assembly molecular distances.191 Fluorescence techniques are ofcourse limited to NM-bioconjugate components that have someform of either intrinsic or extrinsic fluorescence (or fluorescencequenching) capacity; however, given the variety of fluorescentNMs and the vast array of commercially available biomolecular-reactive fluorescent dyes, this should not be considered alimitation.191,192 Obviously if conjugation of a fluorophore tothe NM is desired as part of the final application, then fluores-cence is an excellent method to determine the average number offluorophores per NM, as demonstrated for cellulose and viralNMs.37,89,193 This is especially useful when the NM and a dyehave strong overlapping absorbance. A number of researchershave used the intrinsic fluorescence from tryptophan (Trp)residues, commonly found in protein sequences, to obtaininformation about local changes in tertiary structure upon NMbinding.182,186,187,194 For example, Wu and co-workers lookedat the interaction of β-lactoglobulin with silica NPs and foundthat the extent of protein unfolding, which produced anincrease in the Trp fluorescence, was dependent on thenumber of proteins adsorbed to the NP surface.182 Lower

surface protein concentrations on the NP resulted in greaterprotein unfolding, presumably because the adsorbed proteinhad greater access to free NP surface with which to interact.Rather than relying on Trp emission, Reulen and co-workersused the intrinsic fluorescence of enhanced yellow fluorescentprotein (eYFP) to monitor and confirm its conjugation toliposomal NMs.195

Many biomolecules can be quite easily labeled with extrinsicfluorophores to aid in NM-bioconjugation characterization andoptimization. For example, Hurst and co-workers used fluores-cently labeled DNA to determine the effects of salt concentra-tion, spacer composition, NM size, and sonication on the averageDNA coverage per AuNP,196 while Lockney et al. monitoredfluorescently labeled peptide sequences binding to red clovernecrotic mosaic virus.197 Mittal and co-workers used biotin-avidin affinity to determine the average number of biotin-bindingsites present on SA conjugated QDs, using a biotin-4-fluores-cence (B4F) quenching assay.198 B4F is a commercially availabledye whose fluorescence is quenched upon binding to avidinproteins. While these methods provide information about theaverage biomolecular coverage on the NM surface, distributioninformation is again still lacking. In contrast, single-moleculestudies combined with stepwise photobleaching are increasinglybeing used to obtain ratio distributions.145,146 Delpot and co-workers functionalized 250 nm silica NPs with an ATTO 647Nlabeled DNA 15-mer followed by immobilizing the NP-biocon-jugates onto a cover glass for analysis with confocalmicroscopy.146 After a NP was located in the confocal image,the 633 nm laser was focused on the NP and the fluorescentlylabeled DNA stochastically photobleached. Approximately 100NPs were analyzed by looking at the number of energy levels inthe stepwise decrease of the fluorescence intensity as a functionof time for a particular NP. This was then correlated to a directmeasure of the number of fluorescent DNA on the NP surface.Flow cytometry is occasionally used to characterize biomole-

cules binding toNMs.199,200 Nakamura and co-workers looked atthe binding of green fluorescent protein (GFP) and fluorescein-labeled DNA to epoxy-organosilica,199 while the group of Zhonglooked at poly(lactic-co-glycolic acid) (PLGA) NPs conjugatedto fluorescein labeled Fab0 fragments of anti-HER2 antibody viaEDC/NHS coupling chemistry.200 In both cases, biomoleculebinding resulted in the appearance of a fluorescence signal,confirming the interaction.FRET is a specific fluorescence phenomenon that occurs

between donor and acceptor molecules and is highly dependenton a number of factors, most important of which are the extent ofdonor/acceptor spectral overlap and the separation distancebetween the two. A number of extensive reviews concerningFRET can be found in the literature.191,201 The underlyingprocess has been likened to a molecular ruler with separationsensitivities for donor/acceptor distances proportional to r6 andwhich usually fall in the 1�10 nm range. Medintz and co-workersdemonstrated the unique abilities of semiconductor QDs asdonors in a variety of FRET formats202 and have used thetechnique to establish the orientation of MBP immobilized toa QD (see Figure 9C)203 and to demonstrate that underlyingbioconjugate chemistry can strongly influence the architecture ofQD-DNA conjugates, see below.204 FRET has also been used tomonitor the binding of fluorescently labeled proteins or peptidesto the surface of QDs but is more commonly used as a signaltransduction mechanism in functional biosensing assays.205,206

Morgner and co-workers used the FRET combination of



luminescent terbium complexes (donor) and QDs (acceptor) toinfer information about the shape and size of QD biotin-SAbioconjugates.207 A number of researchers have used AuNPacceptors as quenchers in energy transfer studies with fluorescentdonor species, and the resulting putative surface energy transfer(SET) process observed has been shown to have a nontraditionalr4 distance dependency, allowing in essence the ability to extendthe reach of the molecular ruler.208,209 For example, Sen and co-workers recently used Trp-Au SET to probe conformationalchanges that occur when BSA binds AuNPs.209 Results demon-strated a significant quenching of the BSA Trp (donor) fluores-cence upon protein binding to the AuNP (acceptor) surfaceunder different pH conditions. With the use of SET theory,distances between the donor and acceptor were estimated andrelated to the conformation of the BSA adsorbed to the AuNPsurface.Infrared Spectroscopy. This technique measures the absorp-

tion of IR radiation by a sample resulting from the vibrationalstretching and bending modes within the molecule.172,210 Inparticular, Fourier Transform (FT)-IR is frequently used todemonstrate NM bioconjugation through the appearance ofcharacteristic spectral bands. For example, FT-IR has been usedto characterize lysozyme binding to diamond NPs,180 albuminand peptides to AuNPs,186,211 hemoproteins adsorbed on Au andsilver NPs,119,179 SA bound to silicon NPs,67 β-lactoglobulinadsorbed on silica NPs,182 along with papain and chitosan boundto magnetic NPs.212 In the case of globular proteins, carefulinterpretation of the stretching and bending vibrations in theamide regions, particularly the amide I band (1 600�1 700 cm�1),can provide secondary structural information and hence theconformational state of the bound protein, see Figure 8B.180,182,186

Perevedentseva and co-workers investigated lysozyme adsorbingto different sized nanodiamonds and found that a slight blue shiftin the amide I band was observed only for the 5�50 nmnanodiamonds and not for the larger particles.180 The observedblue shift suggested the adsorbed lysozyme was undergoing aconformational change from a predominately R-helix structureto that of a β-sheet or random coil conformation which sig-nificantly reduced the enzyme activity. Susumu et al. used FT-IRto verify the presence of particular combinations of differentligands that had been cap exchanged onto the QD surface. Theywere able to determine that the QDs were displaying the desiredbiotin moiety along with other molecules by unique amide bandcontribution, see Figure 8C.213

Nuclear Magnetic Resonance and Magnetic ResonanceImaging. Nuclear magnetic resonance (NMR) spectroscopyand magnetic resonance imaging (MRI) measure the intrinsicmagnetic moment of certain nuclei, typically either hydrogen(1H) or carbon (13C), in the presence of an applied magneticfield.172 NMR spectroscopy can provide physical, chemical,structural, or environmental information about the species understudy as well as information concerning the dynamic interactionsof many biological molecules, including proteins and nucleicacids. NMR spectroscopy has been used to characterize humancarbonic anhydrase (HCA) I conformational changes that occurupon interaction with silica,214 alkyne ligation of dendrimers,45

thiol-modified peptides binding AuNPs,211 as well as QD capexchange reactions.64 For example, Pons and co-workers wereable to confirm the highly efficient cap exchange of trioctylpho-sphine/trioctylphosphine oxide (TOP/TOPO) ligands withDHLA ligands on QDs, through the appearance or loss of 1HNMR peaks characteristic of the ligands.64 They were also able toT

able5.

MassSpectroscopy

andTherm

alTechn

iques

technique/types

NM-bioconjugates

analyzed

advantages

disadvantages

refs

massspectroscopy

electrospray

ionizatio

n(ESI),

inductivelycoupledplasma(ICP)

andmatrix

assisted

laserdesorptio

nionizatio

n(M

ALD

I)

viralN

P-biotin,A

uNP-hemoprotein,

silverNP-hemoprotein,T

iO2NP-

dopamine

1.characterizationbasedon

mass-to-chargeratio

.2.

readily

measuresmassincrease

dueto

NM-

biom

oleculebioconjugatio

n.

1.requiresexpensiveequipm

entandexpertise.

2.lim

itedapplicationto

dateforstudying

NM-bioconjugates.

119,215�

218

thermalgravitatio

nalanalysis(T

GA)

magnetic

NP-PE

G,A

uNP-dendrons,

AuN

P-paclitaxel(therapeutic

drug),

AuN

P-PE

G/cyclodextrin

ligands

1.provides

relativeorganicversus

inorganic

compositio

nof

theNM-biomoleculeconjugate.

2.canproviderelativeam

ount

ofthebiom

olecule.

3.provides

thermalstabilityof

theNM-biomolecule

conjugate.

1.inherentlydestructiveto

sample.

2.onlyprovides

bulkinform

ationaboutthe

amount

ofligand.

158,219�

221

calorim

etry

differentialscanningcalorim

etry

(DSC

)andisotherm

altitratio

ncalorim

etry

(ITC)

polymer

NP-proteins,lipidNP-

insulin,polym

erNP-paclitaxel,zinc

NP-BSA

,AuN

P-CTABligands

1.provides

bulkinform

ationabouttheNM-

biom

oleculeconjugate,such

asmeltin

gtemperature,crystallizationandglasstransitio

ninform

ation,

decompositio

nandstabilityinform

ation.

1.onlybulksampleinform

ationprovided.

2.requiresan

appropriatereference.

3.somew

hatlim

itedapplicationto

date.

222�

226

thermophoresis

QD-PEG

1.provides

size

andsurfacepotentialofthe

NM-

biom

oleculeconjugate.

2.canprovideeffectivediam

eter.

1.samplerequiresopticalpropertiesto

aidin

visualization.

2.somew

hatlim

itedapplicationto

date.

34,227,228



confirm the nature of the interaction of the DHLA with the QDsurface through the disappearance of two disparate thiol reso-nances characteristic of the reduced dithiol moiety, which ispresent in the DHLA-only control spectrum but not in theDHLA-capped QDs. This confirmed dithiol affinity and bindingof DHLA for the QD surface. Lundqvist and co-workers per-formed detailed NMR analysis of the human carbonic anhydraseI enzyme before and after interaction with ∼9 nm silica NPs.214

They found that desorption of the protein from the NP surfaceresulted in proteins regaining near-native conformations, withthe most profound differences between native and desorbedenzyme being found in the β-strands typically located in thecenter of the protein structure.Mass Spectroscopy.Mass spectroscopy (MS) comprises the

family of technologies that analyze samples based on their mass-to-charge ratio, Table 5. MS has been used to characterize NM-bioconjugates and has found particular utility in the analysis ofprotein based NMs, such as viral NPs, where mass increases inthe viral coat proteins due to the addition of biotin or fluorophorespecies were successfully monitored using matrix-assisted laserdesorption/ionization (MALDI)-time-of-flight (TOF)-MS.215,216

Through the measured mass increase, the average stochiometryof the additional species added per virus NP could readily bedetermined. MALDI-TOFMS has also been used to qualitativelydemonstrate hemoprotein binding to gold and silver NPs.119

Inductively coupled plasma (ICP) MS was used to determineTiO2 NP binding to a dopamine ligand that was subsequentlyused to complex a gadolinium MRI contrast agent.217 Interest-ingly, QDs and a variety of other NP materials have recentlyfound utility as matrixes for surface-assisted laser desorption/ionization-MS (SALDI-MS) analysis of proteins and peptides.218

Their unique properties allow for sample enrichment and helpovercome high background signals in the low-mass regions.Overall, the application of MS techniques has been fairly limitedfor NM-bioconjugate characterization, and this may in part bedue to the relative cost of the instrumentation, the required levelof expertise needed to run analyses, the destruction of the sampleduring analysis, and the fact that the requisite instruments areusually configured for other analyses, with the latter also usedquite frequently in an automated fashion.Thermal Techniques. As highlighted in Table 5, these

techniques can aid in determining the amount of conjugatedbiomolecule as well as both the NM and the biomolecule’sthermal stability. As with many of the techniques describedabove, sample preparation is the key to successful analysis, wheresample weight, composition, and running conditions (heatingrates, etc.) can all influence the quality of the data obtained.Typically, for techniques like differential scanning calorimetry(DSC), 1�10 mg of sample is required for analysis (dependingon composition); however, newer instruments can now use lowmicrogram amounts. Thermal gravimetric analysis (TGA) is amethod that utilizes a high-precision balance to determinechanges in the weight of a bulk sample relative to changes intemperature and has been used to characterize a variety of NMsfunctionalized with biomolecules, including AuNPs functiona-lized with dendrons, hydroxyapatite grown on Au-fibrin NMs, aswell as the amount of paclitaxel bound to a NM drug deliverysystem.158,219,220 Further calculations can reveal informationabout the average number of ligands attached per NM and theextent of surface functionalization. For example, TGA has beenutilized in conjunction with NMR to approximate the number ofPEG and cyclodextrin ligands on AuNPs.221

DSC and isothermal titration calorimetry (ITC) are otherthermally based methods that can provide bulk informationabout the NM-bioconjugate. DSC is used to study variousmaterial transitions including melting, crystallization, glass tran-sition, and decomposition. Subsequent analysis can indicate thestate of the NM-bioconjugate including the stability of thebiomolecule and structural information on both the NP andbiomolecule including underlying crystallinity and how thedifferent components are interacting with each other. Research-ers have also used DSC to help elucidate the structure andstability of surface coatings of NM-bioconjugates as well as thestate of their therapeutic payloads. For example, DSC has beenused to probe the stability of lipid bilayers when the bilayers wereembedded with silver NPs.222 Sant and co-workers also usedDSC to help elucidate the location of PEG ligands in a poly-(D,L)-lactide NP.223 DSC recrystallization analysis has further beenapplied to determine the stability of solid, lipid NP-insulincomplexes224 and how the individual components of a Zn-NPBSA NM-bioconjugate system interacted with each other.225

ITC provides further potential to investigate the stoichiometry,affinity, and enthalpy of the NM-biomolecule interaction, asdemonstrated by Cedervall and co-workers studying the coronalayering mechanism for various polymeric NMs (based on thecopolymer N-iso-propylacrylamide-co-N-tert-butylacrylamide)and binding proteins (either human plasma or single proteinssuch as HSA or fibrinogen). Overall, as an analytical techniquethis still remains vastly underutilized.226

Thermophoresis or thermodiffusion involves localized heatingof a sample and monitoring the resulting motion of the particlesdue to the temperature gradient.34,227,228 Sperling and co-work-ers compared thermophoresis to various other analytical techni-ques for their ability to determine the size of various PEG-functionalized QDs.34 They found that the reported particlediameters were highly dependent upon the particular techniqueapplied along with the underlying physical principles andassumptions used to determine particle size. In particular, theyfound that while thermophoresis and FCS both measure diffu-sion of dispersed particles to determine effective diameters, thevalues obtained from both techniques differed significantly. Thedifference was speculated to be due to the difference in particleconcentrations used. FCS is a single particle technique requiringdilute particle concentrations, while thermophoresis is an en-semble method requiring higher particle concentrations toensure robust signals. Thermophoresis and its use for the analysisof colloidal suspensions have been recently reviewed in ref 228.

’MODELING

Although not typically considered a “classical” characteriza-tion technique, modeling can provide important insight aboutthe structure and function of NM-bioconjugates. The overall goalof any modeling effort is to gain insight into the properties orbehavior of a system that cannot be directly observed. Formolecular modeling of NM-bioconjugates, available techniquesagain borrow heavily from protein, peptide, and small moleculeresearch and are usually based on either theoretical principlesalone, include empirical or semiempirical parameters, or alter-natively extend to dynamic simulation.

Methods that are derived directly from theoretical principlesand do not include any (semi)empirical parameters are ab initiomethods. The simplest uses classical mechanics to describe thephysical basis behind the models and is referred to as molecular



Figure 9. Continued




mechanics. Here, molecular systems are treated as individualatoms and typically describe them as a point charge with anassociated mass while the interactions between atoms aredescribed by springlike processes; this simplification allows largerassemblies to be considered. Ab initiomethods in which electronsare explicitly considered are termed quantum-chemical methodsand solve the molecular Schr€odinger equation. More complexversions considering electronic structure fall under Hartree�Fock schemes. Semiempirical methods (sometimes based onthe Hartree�Fock scheme) that make approximations andwhich incorporate parameters from empirical data are impor-tant for modeling large systems where other approaches aretoo computationally intensive. Energy minimization is oftenincorporated as lower energy states typically represent morestable configurations and are more likely to represent a largefraction of the molecules in a given environment. Energyminimization is also useful for obtaining a static picture forcomparison to similar systems.

In molecular dynamics simulations, the system configurationis computed over time resulting in atomic trajectories in bothspace and time. This provides information on dynamic processesand can include temperature effects. In cases where only a limited

number of changes to a template structure need to be made, atechnique sometimes referred to as “comparative modeling” canbe effective. This method, for example, builds a three-dimen-sional model for a protein of unknown structure (target) basedon one or more related, known protein structures(templates).229�232 Comparative modeling relies on the similar-ity of target with known template structures. This approach toprotein structure prediction in particular is possible because asmall change in sequence usually results in only a small change inits 3D structure. It is further facilitated by the fact that proteinfolds defining the 3D structure are more highly conserved withina protein family than their primary sequence.233

Important points to consider about NM-bioconjugate model-ing approaches are that even the most complex and computa-tionally intensive modeling methods do not produce an exactsolution; they are all approximations. Further as complexityincreases, more accurate predictions of molecular propertiesmay be obtained but with a greater computational cost. Webriefly survey some examples where different modeling strategieshave potential or already have proven useful to NM-bioconjugatecharacterization.

Figure 9. Modeling NM-bioconjugates: (A) energy-minimized structure of a G3-PAMAM dendrimer conjugate having 32 covalently bound DITC-APEC groups. Terminal adenosine moieties are highlighted in pink for visibility. Overall diameters, ∼80�110 Å, were measured (shown in green)between ligands approximately diagonal to each other through the central core. Reprinted with permission from from ref 235. Copyright 2009 SpringerScienceþBusiness Media. (B) Snapshots of yeast cytochrome c protein structure at the end of an 8 ns MD simulations at 300 and 450 K. Green/blue,N-/C-terminal helices; red, loss of 50%-helicity from the wild-type structure; purple sphere, the AuNP attachment site. Reprinted with permission fromref 59. Copyright 2009National Academy of Sciences.(C) Side view showing the refinedMBP-QDorientational structure with all six dye-label structuresshown in red. Distances from the QD center to each dye position on the protein were determined using FRET, and the refined distances are shown withthe yellow dashed lines. With the use of this refinement,∼45 Å is estimated as the radius of the spheroid (or the distance from the nearest MBP atom ofLys-370, green, in the PDB 1LLS structure to the QD center). This residue in green is the location of the terminal (His)5 sequence which attaches theprotein to the QD. Reprinted with permission from ref 203. Copyright 2004 National Academy of Sciences. (D) Model used to estimate the maximumload of mCherry onto a QD surface. For clarity, the fluorescent groups in each protein molecule are shown in red as space-filling models. On the basis ofthe putative protein contact area and QD surface area, a maximum loading of 18 mCherry per QD was estimated. In the image shown, 11 of the 18protein molecules can be seen. (E) Model of self-assembled QD-mCherry conjugate used to monitor caspase-3 activity. The QD is shown as a spherewith a radius of 30 Å with its DHLA coating as a translucent shell 11.7 Å thick. A single molecule of DHLA is shown in red. The QDHis6-binding regionof the linker peptide is shown in green; the caspase-3 cleavage site is shown in yellow. The β-barrel conformation ofmCherry is depicted as a ribbonwhilethe chromophore is shown as an orange space-filling model within the barrel. Caspase-3 is shown in a stick representation with its active site highlightedin cyan. From this image it is clear that many conformations within the linker peptide with either an extended or bent structure could result in thecleavage site being accessible. (F)Modeling of QD-DNA structures showing a (His)6-peptide-DNA construct bound to 530 nmQDs. The QD is shownas the central blue sphere with a radius of 28 Å. TheDHLA-PEG ligand is indicated by the crimson halo with an estimated extension of 30 Å. DHLA-PEGligands in energy-minimized conformation are shown within the crimson sphere. TheHis6-portion of the peptide is shownwith a yellow ribbon joining itto the DNA. Individual DNA strands within the dsDNA structure are shown in orange and yellow. The rotational extension of the dye molecules areshown by the magenta spheres. An orientation with the DNA and sequential dye placement sites extending linearly outward from the QD surfaceis shown.




Predictive Modeling. Although not fully extended to NM-bioconjugates per se, predictive modeling can clearly aid in thecharacterization and processing of these materials. This wasdemonstrated by Saunders and co-workers who applied a totalinteraction energy model to accurately predict the size/sizedistribution of size-selectively precipitated NPs.234 Selectiveprecipitation of QDs, AuNPs, and other NMs is an importantpart of their initial processing following synthesis into a moredefined sample. In this example, NP precipitation was modeledas a simple two-body interaction which accounted for all relevantinteraction energies. The model had only one manipulatedvariable, which is the amount of antisolvent added to the system.The model was then successfully applied to the size-selectivefractionation of dodecanethiol-stabilized AuNPs dispersed inhexane by CO2 addition/precipitation. An initial sample display-ing a 4.56 ( 1.24 nm size (100%) was selectively precipitatedinto 3 fractions consisting of 6.01( 1.39 nm size (3.6%), 6.20(1.18 nm size (9.2%), and 4.33 ( 1.12 nm size (87.2%).Importantly, these size values are all within 5% of those originallypredicted. This model obviously depends heavily on a prioriknowledge of the physical properties of the NP ligand in relationto the solvent conditions. An important future test will bewhether this model can be extended to a NP dispersed in bufferand stabilized by a biological such as PEG, DNA, or a peptide.This may allow for selection of specifically functionalized materi-als away from reactants or for sorting by ratios of displayedmolecules.Structural Estimates. Modeling is also useful for estimating

somewhat “simpler” parameters such as NP size. Jacobson’sgroup relied on energy minimization to estimate the overall sizeof a functionalized generation 3 polyamidoamine (G3 PAMAM)dendrimer.235 In this example, the author’s functionalized thedendrimer with a chemically reactive nucleoside A2A adenosinereceptor agonist (DITC-APEC) as part of an effort to developmultivalent ligands with enhanced pharmacological effects ascompared to monomeric drugs. The drug was fully incorporatedand displayed on the 32 ligand sites at the dendrimers peripherywith the chemistry utilized. The energy-minimized structure ofthe construct, generated using HyperChem 7.5.2 Amber forcefield, suggested an overall diameter of ∼80�110 Å with anellipsoid shape, see Figure 9A. The author’s were gratified to findthat the DITC-APEC-loaded dendrimers extended the diameterover previously reported derivatives by∼20 Å. It was envisionedthat this could potentially increase the conformational flexibilityof the appended ligands to achieve optimal geometry for efficientbinding to the adenosine receptor.Protein Binding to Nanoparticles. A variety of modeling

strategies have recently helped either elucidate or improve howdifferent proteins attach to NP materials. Mukherjee and co-workers used statistical optimization to improve how alkalineR-amylase immobilized onto supermagnetic iron oxide NPs.236

Plackett�Burman factorial design and response surface metho-dology were utilized to screen the influence of different para-meters such as pH, NP concentration, and the response of theenzyme to the binding process. The authors utilized coefficientsof determination and analysis of variance to validate the pro-posed model along with confirming the size of the NPs by X-raydiffraction and applying FT-IR spectroscopy to confirm enzyme-NP immobilization. Importantly, they found a significant 26-foldincrease in specific activity, improved thermal and storagestability, along with extended reusability of R-amylase afteroptimized binding to the NPs in comparison to that of free

enzyme. These results have important commercial implicationsas this enzyme has significant industrial applications in the paperand brewing industries.Hamad-Schifferli’s group elegantly demonstrated how to

combine recombinant protein engineering, modeling, and othertypes of NP-protein characterization in order to understand howattachment site along with NP material and ligand all influencedsubsequent protein structure.59 As a model system, they utilizedcytochrome c derived from Saccharomyces cerevisiae (Baker’syeast) as its structure has been well studied both experimentallyand computationally. They subsequently conjugated 1.5 nmdiameter negatively charged AuNPs to the protein by introdu-cing site-specific cysteine-thiols into its structure; this allowed forformation of thiol-Au surface linkages. The protein surface sitesutilized for cysteine insertion and hence NP-modification werewidely varied around the cytochrome c structure and alsoincluded areas around the N- and C-termini. Effects of protein-NP labeling were probed with circular dichroism which providedinformation on changes in the proteins helicity. They found thatprotein unfolding was the most severe when the NP labeled nearthe termini as this affected its core folding motif. NP attachmentin the vicinity of charged residues also induced greater structuraldamage and was ascribed to salt-dependent electrostatic inter-actions with the negatively charged NP-surface ligand. Moleculardynamic simulations were then used to both confirm andelucidate the local and global structural perturbations in eachmutant protein upon NP binding, see Figure 9B. Overall thisstudy highlights and again reinforces the importance of judi-ciously choosing a labeling site as part of designing an efficientNP-protein bioconjugate.Modeling of Quantum Dot Bioassemblies. Modeling of

larger assemblies, such as proteins and small molecules as-sembled on the surface of a semiconductor QD, requires anamalgamation of various approaches; it can more easily beconsidered a hybrid derivative of comparative modeling. Tothe greatest extent possible crystallographic structures are used;obtained from either the Protein Data Bank (PDB, www.rcsb.org/pdb) or the Cambridge Structural Database (CSD, www.ccdc.cam.ac.uk). For small molecules or molecular linkers whosestructure is not known, the tools in software such as Chem3D areused to produce energy minimized models. These structures arecombined with experimental results to build models of themolecular assemblies. Conformational parameters of these as-semblies are adjusted to produce a final model which is in closeagreement with the experimental results. Uncertainties in themodel can be teased out by examining alternate conformationsthat place functional groups in extremes of closest approach andfurthest separation.The ability to model these assemblies and have confidence in

the results is derived from initial work with MBP. These werebased on the crystal structure of MBP along with FRETmeasurements made from constructs that placed a fluorescentgroup in several discrete known locations around the proteinsurface to study the QD-MBP complex.203 Each of thesedifferentially labeled MBPs was then separately assembled onthe QD allowing monitoring of each FRET interaction. Theseparation between the QD and fluorophore was estimated fromFRET measurements and the data set of pairwise separationsused to orient the protein on the QD surface. From a separate setof experiments, it was known that a polyhistidine terminus (His5)on the MBP was required for binding to the QD surface. Thevalidity of orientation was confirmed by the proximity of this



sequence to the QD surface, see Figure 9C. Thus other modelstructures could be constructed using the Hisn tail to orientfurther proteins on QDs. Interestingly, this work reflects MBP’sstrong role in prototyping NM-bioconjugates. This, in turn,originates from the excellent understanding of MBP’s structureand function, similar to cytochrome c, above the ability to sitespecifically modify and label it as desired, along with it retainingfunctionality once tethered to a surface or NM.237 Once aputative orientation of a protein on the QD surface is produced,other parameters can be measured from a model. It is alsoimportant to consider when using FRET data that ensembleFRET measurements often reflect an average which can bedominated by those structures having the donor and acceptorin closest proximity.One example of applying this type of modeling to QD

biocomplexes was aimed at determining how many proteinsand peptides canmaximally load onto these nanocrystal materialsas a function of size.238 This understanding has importantimplications for the design of high avidity and efficient sensingQD constructs. For this, models were constructed for QDsdisplaying surfaces of the solubilizing dihydrolipoic acid(DHLA) ligand, a Hisn-appended <20 residue peptide, myoglo-bin, mCherry, and MBP. The contact area or surface footprint of

each molecule on the QD was estimated from these models andused to calculate maximum loadings (Figure 9D). In general,excellent agreement was obtained between the estimated andexperimentally determined protein loadings (i.e., MBP 10/12,mCherry 18/20, myoglobin 24/30 for each model/experimentalresult, respectively). The latter values were determined from gelseparation of increasing ratios of each dye-labeled moleculeexposed to the QDs. It was only in the case of the Hisn-peptidethat any large difference between experimental and estimatedvalues was found. Modeling suggested that a maximum of 140peptides could assemble on the QDs and yet experimentsconsistently yielded average values of only 50 ( 10. Thedifference was ascribed to the presence occupied by the DHLAligand on the QD surface; there were no more available Hisnbinding sites on the surface. Modeling had estimated that from157 to 268 DHLA molecules could attach to the QD surfacedepending upon conformation and, interestingly, other data hadapproximated a value of 140. Overall, this modeling suggestedthat the limiting factor for protein assembly on QDs wasprimarily their own size and how they sterically hinder eachother when assembling on the QD surface, and in contrast, forthe far-smaller Hisn-peptide it was the number of availablebinding sites on the QD.238

Figure 10. Emerging technologies. (A) Image of soluble 60 nm AuNPs captured with the CytoViva using 100� magnification. (B) Nanoparticletracking analysis (NTA) by Nanosight: (i) NTA video frame and (ii) size distribution measurements taken with NTA and DLS of a mixture of 200 and400 nm polystyrene beads (2:1 number ratio). Reprinted with permission from ref 100. Copyright 2010 Springer ScienceþBusiness Media. (C)Scanning ion occlusion sensing by Izon: (i) schematic of the tunable membrane pore that is key to the technology, (ii) detection of various NM-bioconjugates. Blockade duration collected for carboxylated silica particles (light gray, solid curves), SA particles (gray, dashed curves), and λ-DNAparticles (black, dotted curves) at background current levels of 90 nA. Reprinted with permission from ref 250. Copyright 2010 Wiley-VCH VerlagGmbH & Co. KGaA.




In a slightly different configuration, a model of a fluorescentprotein self-assembled to a QD surface by an extended Hisn-linker was constructed to assess whether an enzymatic cleavagesite could be inserted into the linker portion while still allowing atarget protease access to this sequence.239 The point of thisexercise was to evaluate whether this QD-sensor designed tomonitor the activity of the apoptotic effector enzyme caspase-3would not be sterically hindered in silico before recombinantlymodifying the Hisn-linker. With the use of the three-dimensionalcoordinates for mCherry (PDB entry 2H5Q), the enzymecaspase-3 (PDB entry 3EDQ) and a model peptide with stronghomology to the linker, a construct was assembled, see Figure 9E.The model peptide which served as a linker between the QD andmCherry was subsequently energy minimized and assembledusing the structure building tools available in Chimera. A largesubsequence of this linker peptide had been frequently used inthe production and purification of a number of recombinantproteins, and a search of the PDB found this same sequence at theN-terminus of many recombinant proteins which had beencrystallized. Critically, no coordinates were available for thisportion implying that this sequence does not have a well-defined3D structure but is rather present in a random-coil conformation(i.e., many conformations of this peptide are possible). As astarting point for model building of this linking peptide, con-straints were applied to produce an extended conformationallowing estimation of the maximum separation between theQD and the mCherry fluorophore when assembled to thenanocrystal surface. Torsion angles were then adjusted so thatthe His6 region was in contact with the QD surface while the restof the peptide was extended away from this surface. Torsionangles of the DEVD cleavage sequence region were also adjustedto match those of a tetrapeptide inhibitor bound to the active siteof caspase-3. Overall, the model strongly suggested accessibilityto the cleavage site and following the subsequent proteinengineering, a FRET-based sensing ability was indeed experi-mentally confirmed.In a last example, FRET was utilized in combination with

modeling to investigate how attachment chemistry can affectsubsequent QD-DNA conjugate conformation.204 Figure 9Fshows a structural model of one of the three constructs used inthe study, a (His)6-peptide-modified DNA self-assembled onto a530 nm PEGylated QD. The 40 basepair DNA “backbone”sequence was designed to be hybridized with a series of fourcomplementary sequences which allow placement of the accep-tor dye at a series of increasing 10 base pair increments spacedfurther away from the QD donor. In order to estimate QDacceptor distances, the “scope” of the acceptor dye on its linkerwas considered in the modeling and is depicted in the figure as amagenta sphere centered on the attachment point with a radiusof 14.6 Å. With adjustment of the tilt of the DNA moleculerelative to the QD surface, a good correlation could be obtainedbetween measured QD-acceptor distances and those calculatedfrom the model. This same strategy was applied to assembliesusing SA rather than a peptide to attach DNA to a QD. In thatcase, the experimental data in conjunction with the modelingstrongly suggested that the derived FRET distances were anaverage of all possible acceptor positions and reflected theheterogeneous nature of how the SA was originally attached tothe QD.204

’EMERGING TECHNOLOGIES ANDINSTRUMENTATION

As the field of bionanotechnology continues to mature, newtechniques and the requisite instrumentation to measure uniqueaspects of NP-bioconjugates in particular have emerged. It isexpected that these instruments and their capabilities will con-tinue to rapidly evolve over time. Some notable techniques and/or the instrumentation used to perform them are reviewed below.CytoViva with Hyperspectral Imaging. As mentioned in the

Microscopy section, recent advances in optical microscopy haveallowed for sub-100 nm resolution with certain types of opticalmicroscopy systems. One such system, produced by CytoViva,uses a proprietary darkfield-based optical illumination technol-ogy that provides improved contrast and signal-to-noise ratioover traditional optical systems, resulting in sub-100 nm resolu-tion, see Figure 10A. Such resolution can allow tracking of NM-bioconjugates in cells.240When used in conjunction with the dualmode fluorescence module, the user can simultaneously imageboth the fluorescent and nonfluorescent portions of the sample,without the need to switch excitation sources and emission filtersand acquire separate images that are later merged. With theaddition of a second piece of technology produced by thecompany, the Hyperspectral Imaging System, researchers cannot only track the NM-bioconjugates in cells but by utilizing thespectral library capabilities of the instrument are able to positivelyidentify the NM in the cell as well as determine if the NM andbiomolecule are still associated. The hyperspectral imager com-prises a concentric imaging spectrophotometer capable of pro-viding spectral analysis of the sample from 400 to 1000 nm.Typically the user creates a spectral library of NMs of interestand, through a customized software analysis program, spectraloutputs from an unknown sample are compared to the library toaid in identification.Xigo Acorn Area Analysis. Surface area is an important

parameter for monitoring NP reactivity. Traditionally, the sur-face area of NMs is determined through the method developedby Brunauer, Emmit, and Teller (BET). This involves theadsorption of gas molecules on the surface of the particles andtherefore requires that the NM sample be dried. Since dryingmany NM systems can cause aggregation, the surface areadetermined by BET may be lower than the true value of theNMs in solution. The Acorn Area, developed by Xigo Nanotools,uses NMR to measure the wetted surface area of NMs insolution. The technique is based on the principle that liquid incontact with a particle surface has a markedly shorter NMRrelaxation time relative to the bulk liquid. This technique allowsthe possibility of determining surface area pre- and post-NP-bioconjugation in relevant biological solutions.241

Resonance Frequency Devices (Quartz Crystal Microba-lance and Suspended Cantilevers). The use of quartz crystalmicrobalances (QCM) for NM-bioconjugate analysis has be-come more prevalent in the literature recently, especially ascommercial sources for these instruments are expanded. QCMmonitors mass per unit area via the frequency of a quartz crystalresonator.242 The high sensitivity of this method allows QCM tobe used for investigating the interactions between biomoleculesand NMs. For example, QCM allowed monitoring of an anti-body-labeled CoFe2O4/SiO2 NP-functionalized surface duringan immuno-assay for carcinoembryonic antigen.243 Bindingkinetics and the binding specificity of antihuman vascular celladhesion molecule 1 (VCAM-1) functionalized iron oxide NPs



to VCAM-1 modified QCM substrates have also been deter-mined and found to be dependent on the number of antibodiesimmobilized on the NP surface.244 Burg and co-workers demon-strated the use of a suspended cantilever comprising a micro-fluidic channel to weigh single AuNPs that changed theresonance frequency of the cantilever as they transited throughthe device.245 The technique provided femtogram resolution,and the resulting histogram of particle mass could be convertedto size when the density of the material was known.Single Particle Tracking (Nanosight and Others). Nano-

particle tracking analysis (NTA), marketed by laser-illuminatedoptical microscopy to track the light scattering fromNMs that are10 (30)�1000 nm (lower limit depending on reference) andmoving under Brownian motion. Unlike traditional scatteringmethods such as DLS, each scattering source (i.e., NM) is trackedseparately, allowing for individual measurements of particle size(determined using the Stokes�Einstein equation) during popu-lation analysis. In addition, particle concentration (which can bequite challenging to determine when a strong absorbance ormolar extinction coefficient values are not available) may bemeasured with this technique. NTA has also been used to cross-validate DLS data as well as to determine the relative thickness ofprotein opsonized layers on AuNPs.246 A critical evaluation ofNTA versus DLS using polystyrene beads was recently pre-formed by Filipe and co-workers.100 The study highlighted themain advantage of NTA as its unbiased peak resolution ofpolydisperse samples which was not possible with DLS, seeFigure 10B. TheNTA technique currently is more complex to setup and time-consuming to perform but is not influenced by smallamounts of large particles (such as dust) which is problematic forDLS. The NanoSight system is also capable of working underfluorescence mode, and the company recently announced therelease of ζ potential analysis capabilities.Single particle tracking based on fluorescence measurements

rather than scattering has also been demonstrated for sizingfluorescent NMs. Obviously, the particles in this case must beintrinsically fluorescent or extrinsically labeled, and the sizelimitation is governed by the optical resolution of the microscopesystem. Nevertheless, Braeckmans was able to measure 100/200 nm fluorescent nanospheres and liposomes in undilutedwhole blood, something not typically possible with DLS orscattering-based NTA.247 This study used a custom built laserwidefield epi-fluorescence microscope system in combinationwith custom image analysis software to track individual NPs.Subsequent size calculations were determined using a maximumentropy deconvolution method (MEM).Coulter Counter Devices. Coulter counter devices detect

changes in electrical conductance of a small aperture as a sampleis passed through, referred to as the Coulter principle. Its use forcharacterization of NMs has until recently been limited by theability to generate nanosized apertures sensitive to NMs. How-ever, Fraikin and co-workers recently achieved this usingmicrofluidics.248 The microfluidic channel comprised a primarynanoconstriction (for particle detection) and fluidic restrictionregion (which provides a balancing electrical resistance) with asensing electrode located between the two. Particles entering thenanoconstriction via fluid pressure altered the ionic electricalcurrent that, in turn, changed the electrical potential of the fluidin contact with the sensing electrode. Following calibration withknown dilutions of different sized polystyrene NPs, the size andconcentration of T7 bacteriophage was successfully character-ized. In addition analysis of polydisperse samples containing 51,

75, and 117 nm diameter polystyrene beads demonstrated theexcellent resolving capabilities of the technology.Scanning ion occlusion sensing (SIOS) makes use of the

Coulter principle, and Izon of Christchurch, New Zealand, havedeveloped a relatively low-cost commercial instrument based ona tunable nanopore technology that uses SIOS to characterizenanoscale particles, see Figure 10C.249,250 Electrophoretic forceis applied to the nanoscale particle solution causing the particlesto pass through a single, tunable membrane pore, which results ina measurable blockade of the ionic current. Users can varynumerous parameters (pressure, electrophoretic force, and nano-pore size) to determine the particle concentration, electrophore-tic mobility, particle size (single particle resolution), and aggre-gation state/kinetics using a wide range of pH and electrolytebuffers. The system relies on calibration with appropriate nano-scale standards, has a lower limit of detection of 40 nm, and aswith DLS, typically assumes a spherical geometry. That said, theIzon SIOS system has already shown the ability to resolvepolydisperse solutions and can distinguish between DNA mod-ified polystyrene particles and unmodified particles.249,250

’ IMPACT OF NM CHARACTERIZATION ONNANOTOXICOLOGY

Concerns over potential health issues arising from exposure tonanocontaining materials has resulted in the emerging field ofnanotoxicology, which inherently relies on a large characteriza-tion component.24,27,251 Related to this, a number of groups arealready characterizing the NM-protein corona that results whenNMs are introduced into biological environments.226,252 Therelative importance of characterization to this field was reinforcedby a recent editorial in the ACS journal Chemical Research inToxicology, which recommended that nanotoxicology submis-sions include five important features in the manuscript, one ofwhich was “The chemical and physical characteristics of theparticles should be well-defined”.253 There are already in placenumerous toxicity tests available to the researcher, and theapplicability of these tests to the study of NMs and NM-bioconjugates are the subject of a number of excellentreviews.254�256

There are several key points to consider about this field inrelation to characterization. First, results from toxicity tests maynot be meaningful unless the model system under study is wellunderstood; i.e., both the NM and the biological must be wellcharacterized. Second, results may not have validity if thematerials used are contaminated with synthetic byproduct orother biologicals picked up during processing or storage. Con-founding this, recent studies have demonstrated that certainsterilization techniques can be detrimental to the NM, affectingthe stability and physicochemical properties, hence adequatecharacterization may also be essential at this intermediary step aswell.257,258 Samples may also need to be characterized bothbefore and after bioconjugation. A further complication arises asquite often NPs and other NMs will interfere with the working orinterpretation of an in vitro toxicity assay.

Recently, Grainger and Castner highlighted the importance ofsurface contamination and how this can influence toxicitystudies.12 In particular, microbial endotoxins are often over-looked, but given their known ability to trigger inflammatoryresponses, characterization of NMs for the presence of endotoxinis especially important when targeting in vivo applications.259

Endotoxins are lipopolysaccharides found in the outer



membrane of Gram-negative bacteria that are ubiquitous in theenvironment.259 Contamination of NMs with endotoxins duringmanufacture and handling prior to final application is a realconcern for researchers trying to assess their toxic effects,especially the inflammatory responses generated by theirNMs.260,261 This was eloquently highlighted by Vallhov whodemonstrated that endotoxin contamination of their AuNPs wasactually the cause of the cytokine production observed inexposed immature dendritic cells.260 Upon reduction of theendotoxin to below “acceptable” limits, the AuNPs generatedonly minor up-regulation of cytokines. Traditional limulusamebocyte lysate (LAL) assays can be used to determineendotoxin levels; however, as demonstrated recently by Dobro-volskaia, NPs can interfere with these assays and so carefullyselected controls should be included in the studies.256,262 Clearly,to be valid and broadly accepted, NM-bioconjugate toxicityassessments will require carefully designed assays that arescrupulously implemented and appropriately interpreted. Rigorouscharacterization of the NM-bioconjugate will be the startingpoint and foundation for such studies.

’SUMMARY

For NM-bioconjugates to make the leap from an exciting areaof research to reliable products, biomedical materials, andenabling technologies, a number of important questions mustbe addressed in which adequate purification and characterizationwill play a fundamental role. These questions include (1)reproducibility of NM-bioconjugate production, (2) activity ofbiomolecules attached to NM, (3) long-term stability of the NM-bioconjugate, (4) level of control over chemical and biomodifica-tion of the NM surface, (5) purity and potential for contamina-tion of NM-bioconjugate, (6) along with stability throughsterilization and depyrogenation. As highlighted here, a numberof current and emerging analytical techniques are available toaddress these concerns to some extent.

Purification of the NM-bioconjugate is a key issue, andmost ofthe techniques discussed here are at least capable of minimallyremoving unbound biomolecules, although some (e.g., HPLCand electrophoresis) have the intrinsic resolving capability todistinguish different populations of NM-bioconjugates. Unfortu-nately many of these techniques are still only utilized in smallscale purification and emphasis on scaling up the process will berequired to transition into the commercial sector. Purificationwould likewise benefit from improved bioconjugation chemis-tries that limit NM aggregation and reduce the polydispersity ofthe products. Purity concerns also arise from the in situ and ex situenvironment in which the NM-bioconjugate is synthesized orused, especially given the endotoxin contamination discussedabove. These concerns can be somewhat alleviated throughcareful control of the reagents and environment used duringthe synthesis, purification, storage, and subsequent use of thefinal NM-bioconjugate.

Final characterization of the purified NM-bioconjugates willultimately require a combination of techniques to fully address allphysicochemical characteristics, in addition to the bioconjugationmetrics of interest. Not all the techniques discussed here areapplicable to all types of NMs or their subsequent NM-bioconju-gates. In addition many of the techniques require some form ofsample manipulation prior to analysis, such as drying or suspen-sion in ultrapure liquids. This manipulation may result in non-physiological states and perturbed NM-bioconjugate properties,

and hence interpretation should be erred on the side of caution.The key issues to these technologies being incorporated intoroutine laboratory practice by researchers will be relative cost, easeof use, resolution capabilities, sample preparation requirements,ease of data interpretation, versatility, bulk versus single particleanalysis, etc. Some techniques such as chromatography andelectrophoresis are already routinely used, are relatively cheap,are widely available, can readily confirm biomolecular attachmentto the NM surface, and provide either/both purification andcharacterization information on the bioconjugate product. DLSand ζ potential characterization of NM-bioconjugates are againrelatively cheap and simple to perform providing hydrodynamicradius, aggregation state, and surface potential information. SEMand TEM are mainly used for characterization of the NM itself(not so much the biomolecule to date) and are relatively moreexpensive, equipment and maintenance-wise; however, they char-acterize the size and shape of the NM on an individual particlebasis. AFM in contrast can divulge a range of information aboutboth the NM and the biomolecule again on a single-particle basis.Many of the spectroscopic techniques (which range in cost)provide bulk analysis of the NM-bioconjugate, with NMR andIR spectroscopy demonstrating the ability to characterize biomo-lecular conformational states on the NM. High-resolution micro-scopy and a number of the emerging technologies discussedprovide exciting alternatives for characterizing particular metricsof interest. It is anticipated that many of the requisite instrumentsfor these analyses will soon start to populate core facilities andnational user centers. Another area that will benefit NM-charac-terization as awhole is the development and availability of standardreference materials for calibrating many of the techniques de-scribed here. The U.S. National Institute of Standards andTechnology (NIST, www.nist.gov/nanotechnology-portal.cfm)continues to develop a variety of NM standard reference materials,including gold, silica, polystyrene NPs, and single-walled carbonnanotubes (SWCNTs), that provide both size and shape variation.However, given the drive to develop multifunctional NMs,standards that reflect ligand chemistry and/or bioconjugation willalso soon need to be developed.

Both the techniques and instruments described here arecontinually evolving to meet the demands of nanoscale char-acterization and while bulk analysis will continue to play asignificant role, additional focus is now being placed on techni-ques capable of purifying and characterizing individual NP-bioconjugate populations. Such single particle techniques do,however, need to ensure enough discrete samples are analyzed toobtain statistically relevant data that reflect the underlyingproperties of the ensemble sample population. Importantly,judicious interpretation of all results in the correct context shouldbe another primary concern. For example, spectroscopy maysuggest a particular NM-biological interaction in solution; however,this may not necessarily translate into a strong binding event thatcan survive chromatographic separation to yield a viable bioconju-gate for a given application. In summary, many of the technologiesdescribed here will play a pivotal role in the development of noveland increasingly complex NP-bioconjugates and these, in turn, willbe indispensable to the future of bionanotechnology.

’AUTHOR INFORMATION

Corresponding Author*E-mail: [email protected] (K.E.S.); [email protected] (I.L.M.).



’BIOGRAPHIES

Dr. Kim E. Sapsford studied chemistry at the University ofEast Anglia (UEA, Norwich, U.K.) and in 2001 received her Ph.D. in analytical chemistry developing optical biosensors. In 2001, shemoved to the Center for Bio/Molecular Science and Engineering atthe U.S. Naval Research Laboratory, where she worked (until 2007)on creating fluorescent-based biosensors using the Array Biosensortechnology developed by Dr. Frances Ligler. Currently she is a StaffFellow at the U.S. Food and Drug Administration (FDA) in theOffice of Science and Engineering Laboratories (OSEL), Division ofBiology (DB). Her work involves assessing biotechnology concern-ing public health safety in particular future biosensing technologies.

Dr. Katherine M. Tyner studied chemistry at Carleton College(Minnesota). In 2004, she received her Ph.D. in materialschemistry from Cornell University (New York) under ProfessorEmmanuel Giannelis where she studied nanoparticles for geneand drug delivery applications. In 2004, she moved to theUniversity of Michigan, where she completed a postdoctoralfellowship in nanoparticle sensors under the direction of Pro-fessors Raoul Kopelman and Martin Philbert. Currently, Katherineis a chemist at the U.S. Food and Drug Administration (FDA) inthe Center for Drug Evaluation and Research (CDER), Office ofTesting and Research (OTR), Division of Drug Safety Research(DDSR). Her work involves assessing nanotechnology as itrelates to the safety and efficacy of therapeutics.

Dr. Benita J. Dair is a research materials engineer with a B.S.from Cornell and Ph.D. from MIT. She has experience inmaterials characterization, including electron microscopy, lightscattering, and imaging techniques. She joined FDA’s Center forDevices and Radiological Health (CDRH), Office of Science andEngineering laboratories (OSEL) in 2004 and currently serves asthe Deputy Director of the Division of Chemistry and MaterialsScience (DCMS). Her interests in nanomaterials are character-ization of their properties and assessing their interactions withbiological systems to ensure safety and efficacy of medical devicesincorporating nanotechnologies.

Dr. Jeffrey R. Deschamps has worked at the Naval ResearchLaboratory since 1985 on structural studies, structure functionrelationships of biological molecules, and biosensors. Prior to hisposition at NRL, Dr. Deschamps was a postdoctoral fellow in theDepartment of Pharmacology at the Johns Hopkins School ofMedicine.He is the author of over 190 publications in diverse areassuch as energetic materials, peptide and protein structure, andbiosensor design and evaluation. He was elected to the AmericanCrystallographic Association’s “Data, Standards, and ComputingCommittee”. His training and experience in biochemistry, struc-tural studies, and molecular modeling puts him in a uniqueposition to model inorganic-biomolecular composites. His recentwork with others at NRL on nanomaterials has resulted in widelycited new methods for characterizing these complex assemblies.

Dr. Igor L. Medintz studied chemistry and forensic scienceat John Jay College of Criminal Justice, City University ofNew York (CUNY). In 1998, he received his Ph.D. in molecularbiology under Prof. Corinne Michels of Queens College (alsoCUNY). He carried out postdoctoral research under Prof.Richard A. Mathies (UC Berkeley) on the development ofFRET-based genetic assays using microfabricated devices. Since2004 he has been employed as a Research Biologist at the Centerfor Bio/Molecular Science and Engineering of the U.S. NavalResearch Laboratory in Washington D.C. His current researchinvolves developing chemistries to controllably bridge the

biological�nanomaterial interface with a focus on the designof new biosensing hybrids that incorporate energy transfer.

’ACKNOWLEDGMENT

I.L.M. and K.E.S thank F. Ligler at NRL for initial encourage-ment. I.L.M. acknowledges the NRL-NSI, ONR, DTRA, andDARPA for financial support. K.E.S would like to thank Dr. T.Umbreit and Dr. B. Casey (FDA) for their helpful commentsduring the preparation of this manuscript. This paper reflects thecurrent thinking and experience of the authors. No officialsupport or endorsement by the U.S. Food and Drug Adminis-tration is intended or should be inferred.

’REFERENCES

(1) Rana, S.; Yeh, Y.-C.; Rotello, V. M. Curr. Opin. Chem. Biol. 2010,14, 828–834.

(2) Willner, I.; Willner, B. Nano Lett. 2010, 10, 3805–3815.(3) Agasti, S. S.; Rana, S.; Park, M.-H.; Kim, C. K.; You, C.-C.;

Rotello, V. M. Adv. Drug Delivery Rev. 2010, 62, 316–328.(4) Aubin-Tam, M.-E; Hamad-Schifferli, K. Biomed. Mater. 2008,

3, 034001.(5) Thanh, N. T. K.; Green, L. A. W.Nano Today 2010, 5, 213–230.(6) Kim, B. Y. S.; Rutka, J. T.; Chan, W. C. W. N. Engl. J. Med. 2010,

363, 2434–2443.(7) EU Scientific Committee on Emerging and Newly Identified

Health Risks (SCENIHR). Report: Scientific basis for the definition of theterm “Nanomaterial”, 2010, http://ec.europa.eu/health/scientific_com-mittees/emerging/docs/scenihro030.pdf (accessed March 24, 2011).

(8) Rosenthal, S. J.; Chang, J. C.; Kovtun, O.; McBride, J. R.;Tomlinson, I. D. Chem. Biol. 2011, 18, 10–24.

(9) Figuerola, A.; Di Corato, R.;Manna, L.; Pellegrino, T. Pharmacol.Res. 2010, 62, 126–143.

(10) Schnorr, J. M.; Swager, T. M. Chem. Mater. 2011, 23, 646–657.(11) Subbiah, R.; Veerapandian, M.; Yun, K. S. Curr. Med. Chem.

2010, 17, 4559–4577.(12) Grainger, D. W.; Castner, D. G. Adv. Mater. 2008, 20, 867–877.(13) Gil, P. R.; Oberdorster, G.; Elder, A.; Puntes, V. Z.; Parak, W. L.

ACS Nano 2010, 10, 5527–5531.(14) Baer, D. R.; Gaspar, D. J.; Nachimuthu, P.; Techane, S. D.;

Casner, D. G. Anal. Bioanal. Chem. 2010, 396, 983–1002.(15) Gaumet, M.; Vargas, A.; Gurny, R.; Delie, F. Eur. J. Pharm.

Biopharm. 2008, 69, 1–9.(16) Kranz, C.; Eaton, D. C.; Mizaikoff, B. Anal. Bioanal. Chem.

2011, 399, 2309–2311.(17) Medintz, I. L.; Uyeda, H. T.; Goldman, E. R.;Mattoussi, H.Nat.

Mater. 2005, 4, 435–446.(18) Hermanson, G. T. Bioconjugate Techniques, 2nd ed.; Academic

Press: Amsterdam, The Netherlands, 2008.(19) Perrier, T.; Saulnier, P.; Benoit, J.-P. Chem.—Eur. J. 2010,

16, 11516–11529.(20) Nguyen, D. T.; Kim, D.-J.; Kim, K.-S.Micron 2011, 42, 207–227.(21) Praig, V. G.; Hall, E. A. H.Anal. Chim. Acta 2003, 500, 323–336.(22) Mullen, D. G.; Desai, A. M.; Waddell, J. N.; Cheng, X.; Kelly,

C. V.; McNerny, D. Q.; Majoros, I. J.; Baker, J. R.; Sander, L. M.; Orr,B. G.; Banaszak Holl, M. M. Bioconjugate Chem. 2008, 19, 1748–1752.

(23) Powers, K. W.; Palazuelos, M.; Moudgil, B. M.; Roberts, S. M.Nanotoxicology 2007, 1, 42–51.

(24) Warheit, D. B. Nano Lett. 2010, 10, 4777–4782.(25) Murray, R. Anal. Chem. 2009, 81, 1723.(26) Boverhof, D. R.; David, R. M. Anal. Bioanal. Chem. 2010,

396, 953–961.(27) Krug, H. F.; Wick, P. Angew. Chem., Int. Ed. 2011, 50, 2–21.(28) McNeil, S. E.Characterization of Nanoparticles Intended for Drug

Delivery; Methods in Molecular Biology, Vol. 697; Humana Press:Totowa, NJ, 2010.



(29) Tyner, K. M.; Nakissa, S.Methods Mol. Biol. 2011, 697, 17–31.(30) Dair, B. J.; Tyner, K.; Sapsford, K. E.Methods in Bioengineering:

Nanoscale Bioengineering and Nanomedicine; Artech House: Boston, MA,2009.(31) Liu, H.; Webster, T. J. Biomaterials 2007, 28, 354–369.(32) Hassell€ov, M.; Readman, J. W.; Ranville, J. F.; Tiede, K.

Ecotoxicology 2008, 17, 344–361.(33) Li, L.; Mu, Q.; Zhang, B.; Yan, B. Analyst 2010, 135, 1519–1530.(34) Sperling, R. A.; Liedl, T.; Duhr, S.; Kudera, S.; Zanella, M.; Lin,

C.-A. J.; Chang, W. H.; Braun, D.; Parak, W. J. J. Phys. Chem. C 2007,111, 11552–11559.(35) Medley, C. D.; Bamrungsap, S.; Tan, W.; Smith, J. E. Anal.

Chem. 2011, 83, 727–734.(36) Goldman, E. R.; Anderson, G. P.; Tran, P. T.; Mattoussi, H.;

Charles, P. T.; Mauro, J. M. Anal. Chem. 2002, 74, 841–847.(37) Sapsford, K. E.; Soto, C. M.; Blum, A. S.; Chatterji, A.; Lin, T.;

Johnson, J. E.; Ligler, F. S.; Ratna, B. R. Biosen. Bioelectron. 2006,21, 1668–1673.(38) Hoshino, Y.; Haberaecker, W. W., III; Kodama, T.; Zeng, Z.;

Okahata, Y.; Shea, K. J. J. Am. Chem. Soc. 2010, 132, 13648–13650.(39) Trapiella-Alfonso, L.; Montoro Bustos, A. R.; Encinar, J. R.;

Costa-Fern�andez, J. M.; Pereiro, R.; Sanz-Medel, A. Nanoscale 2011,3, 954–957.(40) Eck, W.; Craig, G.; Sigdel, A.; Ritter, G.; Old, L. L.; Tang, L.;

Brennan, M. F.; Allen, P. J.; Mason, M. D. ACS Nano 2008,2, 2263–2772.(41) Otto, D. P.; Vosloo, B. C. M.; de Villiers, M. M. J. Liq.

Chromatogr. Relat. Technol. 2007, 30, 2489–2514.(42) Claridge, S. A.; Liang, H. W.; Basu, S. R.; Fr�echet, J. M. J.;

Alivisatos, A. P. Nano Lett. 2008, 8, 1202–1206.(43) Wu, C.; Schneider, T.; Zeigler, M.; Yu, J.; Schiro, P. G.;

Burnham, D. R.; McNeill, J. D.; Chiu, D. T. J. Am. Chem. Soc. 2010,132, 15410–15417.(44) Kuo, Y.-C.; Wang, Q.; Ruengruglikit, C; Yu, H.; Huang, Q.

J. Phys. Chem. C 2008, 112, 4818–4824.(45) Mullen, D. G.; Fang, M.; Desai, A.; Baker, J. R.; Orr, B. G.;

Banaszak Holl, M. M. ACS Nano 2010, 4, 657–670.(46) Stavis, S. M.; Geist, J.; Gaitan, M. Lab Chip 2010,

10, 2618–2621.(47) Schimpf, M., Caldwell, K. D., Giddings, J. C., Eds. Field Flow

Fractionation Handbook; Wiley Interscience: New York, 2000.(48) Williams, S. K. R.; Runyon, J. R.; Ashames, A. A. Anal. Chem.

2011, 83, 634–642.(49) Roda, B.; Zattoni, A.; Reschiglian, P.; Moon, M. H.; Mirasoli,

M.; Michelini, E.; Roda, A. Anal. Chim. Acta 2009, 635, 132–143.(50) Cho, T. J.; Hackley, V. A. Anal. Bioanal. Chem 2010,

398, 2003–2018.(51) Ma, P. L.; Buschmann, M. D.; Winnik, F. M. Biomacromolecules

2010, 11, 549–554.(52) Surugau, N.; Urban, P. L. J. Sep. Sci. 2009, 32, 1889–1906.(53) Pyell, U. Electrophoresis 2010, 31, 814–831.(54) Pellegrino, T.; Sperling, R. A.; Alivisatoes, A. P.; Parak, W. J.

J. Biomed. Biotechnol. 2007, 26796.(55) Park, S.; Hamad-Schifferli, K. J. Phys. Chem. C 2008,

112, 7611–7616.(56) Berti, L.; D’Agostino, P. S.; Boeneman, K.; Medintz, I. L. Nano

Res. 2009, 2, 121–129.(57) Doane, T. L.; Cheng, Y.; Babar, A.; Hill, R. J.; Burda, C. J. Am.

Chem. Soc. 2010, 132, 15624–15631.(58) B€ucking, W.;Massadeh, S.; Merkulov, A.; Xu, S.; Nann, T.Anal.

Bioanal. Chem. 2010, 396, 1087–1094.(59) Aubin-Tam, M.-E.; Hwang, W.; Hamad-Schifferli, K. Proc. Natl.

Acad. Sci. U.S.A. 2009, 106, 4095–4100.(60) Quarta, A.; Ragusa, A.; Deka, S.; Tortiglione, C.; Tino, A.;

Cingolani, R.; Pellegrino, T. Langmuir 2009, 25, 12614–12622.(61) Steinmetz, N. F.; Mertens, M. E.; Taurog, R. E.; Johnson, J. E.;

Commandeur, U.; Fischer, R.; Manchester, M. Nano Lett. 2010,10, 305–312.

(62) Mastroianni, A. J.; Claridge, S. A.; Alivisatos, A. P. J. Am. Chem.Soc. 2009, 131, 8455–8459.

(63) Dif, A.; Boulmedais, F.; Pinot, M.; Roullier, V.; Baudy-Floc’h,M.; Coquelle, F. M.; Clarke, S.; Neveu, P.; Vignaux, F.; Le Borgne, R.;Dahan, M.; Gueroui, Z.; Marchi-Artzner, V. J. Am. Chem. Soc. 2009,131, 14738–14746.

(64) Pons, T.; Uyeda, H. T.; Medintz, I. L.; Mattoussi, H. J. Phys.Chem. B 2006, 110, 20308–20316.

(65) Liu, H. Y.; Gao, X. Bioconjugate Chem. 2011, 22, 510–517.(66) Huang, X.; Weng, J.; Sang, F.; Song, X.; Cao, C.; Ren, J.

J. Chromatogr., A 2006, 1113, 251–254.(67) Choi, J.; Wang, N. S.; Reipa, V. Bioconjugate Chem. 2008,

19, 680–685.(68) Helle, A.; Hirsj€arvi, S.; Peltonen, L.; Hirvonen, J.; Wiedmer,

S. K. J. Chromatogr., A 2008, 1178, 248–255.(69) Suntornsuk, L. Anal. Bioanal. Chem. 2010, 398, 29–52.(70) Gahnon, Z.; Senapati, S.; Chang, H.-C. Electrophoresis 2010,

31, 666–671.(71) Vahey, M. D.; Voldman, J. Anal. Chem. 2009, 81, 2446–2455.(72) Ribeiro, C.; Arizaga, G. G. C.; Wypych, F.; Sierakowski, M.-R.

Int. J. Pharm. 2009, 367, 204–210.(73) Liu, Y.; Franzen, S. Bioconjugate Chem. 2008, 19, 1009–1016.(74) Saini, V.; Martyshkin, D. V.; Mirov, S. B.; Perez, A.; Perkins, G.;

Ellismann, M. H.; Towner, V. D.; Wu, H.; Pereboeva, L.; Borovjagin, A.;Curiel, D. T.; Everts, M. Small 2008, 2, 262–269.

(75) Bonaccorso, F.; Hasan, T.; Tan, P. H.; Sciascia, C.; Privitera, G.;Di Marco, G.; Gucciardi, P. G.; Ferrari, A. C. J. Phys. Chem. C 2010,114, 17267–17285.

(76) Jamison, J. A.; Krueger, K. M.; Yavuz, C. T.; Mayo, J. T.;LeCrone, D.; Redden, J. J.; Colvin, V. L. ACS Nano 2008, 2, 311–319.

(77) Jamison, J. A.; Kruger, K. M.; Mayo, J. T.; Yavuz, C. T.; Redden,J. J.; Colvin, V. L. Nanotechnology 2009, 20, 355702.

(78) Planken, K. L.; C€olfen, H. Nanoscale 2010, 2, 1849–1869.(79) Tiwari, D. K.; Tanaka, S.-I.; Inouye, Y.; Yashizawa, K.; Watanabe,

T. M.; Jin, T. Sensors 2009, 9, 9332–9354.(80) Vannoy, C. H.; Xu, J.; Leblanc, R. M. J. Phys. Chem. C 2010,

114, 766–773.(81) Trefry, J. C.; Monahan, J. L.; Weaver, K. M.; Meyerhoefer, A. J.;

Markopolous, M. M.; Arnold, Z. S.; Wooley, D. P.; Pavel, I. E. J. Am.Chem. Soc. 2010, 132, 10970–10972.

(82) Gaborski, T.R.; Snyder, J. L.; Striemer,C.C.; Fang,D.Z.;Hoffman,M.; Fauchet, P. M.; McGrath, J. L. ACS Nano 2010, 4, 6973–6981.

(83) Zhang, Y.; Schnoes, A. M.; Clapp, A. R. Appl. Mater. Interfaces2010, 2, 3384–3395.

(84) Scherer, C.; Noskov, S.; Utech, S.; Bantz, C.; Mueller, W.;Krohne, K.; Maskos J. Nanosci. Nanotechnol. 2010, 10, 6834–6839.

(85) Rameshwar, T.; Samal, S.; Lee, S.; Kim, S.; Chi, J.; Kim, I. S.J. Nanosci. Nanotechnol. 2006, 6, 2461–2467.

(86) Schmidt, B.; Loeschner, K.; Hadrup, N.; Mortensen, A.; Sloth,J. J.; Koch, C. B.; Larsen, E. H. Anal. Chem. 2011, 83, 2461–2468.

(87) Contado, C.; Dalpiaz, A.; Leo, E.; Zborowski, M.; Williams,P. S. J. Chromatogr., A 2007, 1157, 321–335.

(88) Contado, C.; Argazzi, R. J. Chromatogr., A 2009,1216, 9088–9098.

(89) Park, S.; Sinha, N.; Hamad-Schifferli, K. Langmuir 2010,26, 13071–13075.

(90) Robertson, K. L.; Soto, C. M.; Archer, M. J.; Odoemene, O.;Liu, J. L. Bioconjugate Chem. 2011, 22, 595�604.

(91) Langer, K.; Anhorn, M. G.; Steinhauser, I.; Dreis, S.; Celebi, D.;Schrickel, N.; Faust, S.; Vogel, V. Int. J. Pharm. 2008, 347, 109–117.

(92) Calabretta, M.; Jamison, J. A.; Falkner, J. C.; Liu, Y.; Yuhas,B. D.; Mathews, K. S.; Colvin, V. L. Nano Lett. 2005, 5, 963–967.

(93) Lees, E. M.; Gunzburg, M. J.; Nguyen, T. L.; Howlett, G. J.;Rothacker, J.; Nice, E. C.; Clayton, A. H. A.; Mulvaney, P. Nano Lett.2008, 8, 2883–2890.

(94) Lundqvist, M.; Sethson, I.; Jonsson, B.-H. Langmuir 2004,20, 10639–10647.



(95) Falabella, J. B.; Cho, T. J.; Ripple, D. C.; Hackley, V. A.; Tarlov,M. J. Langmuir 2010, 26, 12740–12747.(96) Pease, L. F., III; Feldblyum, J. I.; DePaoli Lacaerda, S. H.; Liu,

Y.; Hight Walker, A. R.; Anumolu, R.; Yim, P. B.; Clarke, M. L.; Kang,H. G.; Hwang, J. ACS Nano 2010, 4, 6982–6988.(97) Takahashi, K.; Kato, H.; Saito, T.; Matsuyama, S.; Kinugasa, S.

Part. Part. Syst. Charact. 2008, 25, 31–38.(98) Jans, H.; Liu, X.; Austin, L.; Maes, G.; Huo, Q. Anal. Chem.

2009, 81, 9425–9432.(99) Murdock, R. C.; Stole-Braydich, L.; Schrand, A. M.; Schlager,

J. J.; Hussain, S. M. Toxicol. Sci. 2008, 101, 239–253.(100) Filipe, V.; Hawe, A.; Jiskoot, W. Pharm. Res. 2010, 27, 796–810.(101) Khlebtsov, N. G.; Bogatyrev, V. A.; Khlebtsov, B. N.; Dykman,

L. A.; Englebienne, P. Colloid J. 2003, 65, 622–635.(102) Xing, Y.; So, M. K.; Koh, A. L.; Sinclair, R.; Rao, J. Biochem.

Biophys. Res. Commun. 2008, 372, 388–394.(103) Ipe, B. I.; Shukla, A.; Lu, H.; Zou, B.; Rehage, H.; Niemeyer,

C. M. Chem. Phys. Chem. 2006, 7, 1112–1118.(104) Wu, B.; Chen, Y.; M€uller, J. D. Biophys. J. 2008, 94, 2800–2808.(105) M€uller, C. B.; Loman, A.; Richtering, W.; Enderlein, J. J. Phys.

Chem. B 2008, 112, 8236–8240.(106) Roy, S.; Dixit, C. K.; Woolley, R.; MacCraith, B. D.; O’Kennedy,

R.; McDonagh, C. Langmuir 2010, 26, 18125–18134.(107) Rusu, L.; Gambhir, A.; McLaughlin, S.; R€adler, J. Biophys. J.

2004, 87, 1044–1053.(108) Ray, K.; Zhang, J.; Lakowicz, J. R. Anal. Chem. 2008,

80, 7313–7318.(109) Tang, L.; Dong, C.; Ren, J. Talanta 2010, 81, 1560–1567.(110) Ruan, Q.; Tetin, S. Y. Anal. Biochem. 2008, 374, 182–195.(111) Pons, T.; Medintz, I. L.; Wang, X.; English, D. S.; Mattoussi,

H. J. Am. Chem. Soc. 2006, 128, 15324–15331.(112) Qiu, X.; Tang, L. C.; Dong, C. Q.; Ren, J. C. Chin. Chem. Lett.

2010, 21, 1227–1230.(113) Bergman, L.; Rosenholm, J.; €Ost, A.-B.; Duchanoy, A.;

Kankaanp€a€a, P.; Heino, J.; Lind�en,M. J. Nanomater. 2008, 2008, 712514.(114) Gomes, I.; Santos, N. C.; Oliveira, M. A.; Quintas, A.; Eaton,

P.; Pereira, E.; Franco, R. J. Phys. Chem. C 2008, 112, 16340–16347.(115) Liu, Z.; Li, X.; Tabakman, S. M.; Jiang, K.; Fan, S.; Dai, H.

J. Am. Chem. Soc. 2008, 130, 13540–13541.(116) Yang, Q.; Shuai, L.; Zhou, J.; Lu, F.; Pan, X. J. Phys. Chem. B

2008, 112, 12934–12939.(117) Mannelli, I.; Marco, M.-P. Anal. Bioanal. Chem. 2010,

398, 2451–2469.(118) Kneipp, J.; Kneipp, H.; Wittig, B.; Kneipp, K. Nanomed.:

Nanotechnol., Biol. Med. 2010, 6, 214–226.(119) Tom, R. T.; Samal, A. K.; Sreeprasad, T. S.; Pradeep, T.

Langmuir 2007, 23, 1320–1325.(120) Zhang, D.; Neumann, O.; Wang, H.; Yuwono, V. M.; Barhou-

mi, A.; Perham, M.; Hartgerink, J. D.; Wittung-Stafshede, P.; Halas, N. J.Nano Lett. 2009, 9, 666–671.(121) Fu, A.; Zhou, N.; Huang, W.; Wang, D.; Zhang, L.; Li, H.

J. Biomed. Mater. Res., Part A 2005, 74A, 156–163.(122) Brunner, T. J.; Bohner, M.; Dora, C.; Gerber, C.; Stark, W. J.

J. Biomed. Mater. Res., Part B: Appl. Biomater. 2008, 84B, 350–362.(123) Tyner, K. M.; Roberson, M. S.; Berghorn, K. A.; Li, L.;

Gilmour, R. F., Jr.; Batt, C. A.; Giannelis, E. P. J. Controlled Release2004, 100, 399–409.(124) Wang, X.; Du, Y.; Luo, J. Nanotechnology 2008, 19, 065707.(125) Tyner, K. M.; Schiffman, S. R.; Giannelis, E. P. J. Controlled

Release 2004, 95, 501–514.(126) Choy, J. H.; Kwak, S. Y.; Jeong, Y. J.; Park, J. S. Angew. Chem.,

Int. Ed. 2000, 39, 4041–4045.(127) Jenning, V.; Schafer-Korting,M.; Gohla, S. J. Controlled Release

2000, 66, 115–126.(128) Castelletto, V.; Newby, G. E.; Zhu, Z.; Hamley, I. W. Langmuir

2010, 26, 9986–9996.(129) Lipfert, J.; Doniach, S.Annu. Rev. Biophys. Biomol. Struct. 2007,

36, 307–237.

(130) Hammouda, B. Polymer Rev. 2010, 50, 14–39.(131) Park, S. Y.; Lytton-Jean, A. K. R.; Lee, B.; Weigand, S.; Schatz,

G. C.; Mirkin, C. A. Nature 2008, 451, 553–556.(132) Chen, M.-C.; Wong, H.-S.; Lin, K.-J.; Chen, H.-L.; Wey, S.-P.;

Sonaje, K.; Lin, Y.-H.; Chu, C.-Y.; Sung, H.-W. Biomaterials 2009,30, 6629–6637.

(133) Bhattacharyya, M. S.; Hiwale, P.; Piras, M.; Medda, L.; Steri,D.; Piludu, M.; Salis, A.; Monduzzi, M. J. Phys. Chem. C 2010,114, 19928–19934.

(134) McKenzie, L. C.; Haben, P. M.; Kevan, S. D.; Hutchison, J. E.J. Phys. Chem. C 2010, 114, 22055–22063.

(135) Serefoglou, E.; Oberdisse, J.; Staikos, G. Biomacromolecules2007, 8, 1195–1199.

(136) Paul, A.; Vicent, M. J.; Duncan, R. Biomacromolecules 2007,8, 1573–1579.

(137) Jin, S.-E.; Bae, J. W.; Hong, S. Microsc. Res. Tech. 2010,73, 813–823.

(138) Midgley, P. A.; Durkan, C. Mater. Today 2008, 11, 8–11.(139) Dedecker, P.; Hofkens, J.; Hotta, J. Mater. Today 2008,

11, 12–21.(140) Hell, S. W. Nat. Methods 2009, 6, 24–32.(141) Wenger, J.; Rigneault, H. Int. J. Mol. Sci. 2010, 11, 206–221.(142) Wang, Z.; Guo, W.; Li, L.; Luk’yanchuk, B.; Khan, A.; Liu, Z.;

Chen, Z.; Hong, M. Nat. Commun. 2011, 2, 218.(143) Pons, T.;Mattoussi, H.Ann. Biomed. Eng. 2009, 37, 1934–1959.(144) Zhang, J.; Fu, Y.; Chowdhury, M. H.; Lakowicz, J. R. J. Phys.

Chem. C 2008, 112, 18–26.(145) Casanova, D.; Gaiume, D.; Moreau, M.; Martin, J.-L.; Gacoin,

T.; Boilot, J.-P.; Alexandrou, A. J. Am. Chem. Soc. 2007,129, 12592–12593.

(146) Delport, F.; Deres, A.; Hotta, J.; Pollet, J.; Verbruggen, B.; Sels,B.; Hofkens, J.; Lammertyn, J. Langmuir 2010, 26, 1594–1597.

(147) Cobley, C. M.; Skrabalak, S. E.; Campbell, D. J.; Xia, Y.Plasmonics 2009, 4, 171–179.

(148) Chorny, M.; Fishbein, I.; Alferiev, I. S.; Nyanguile, O.; Gaster,R.; Levy, R. J. Mol. Ther. 2006, 14, 382–391.

(149) Jensen, G. C.; Yu, X.; Gong, J. D.; Munge, B.; Bhirde, A.; Kim,S. N.; Papadimitrakopoulos, F.; Rusling, J. F. J. Nanosci. Nanotechnol.2009, 9, 249–255.

(150) Thobhani, S.; Attree, S.; Boyd, R.; Kumarswami, N.; Noble, J.;Szymanski, M.; Porter, R. A. J. Immunol. Methods 2010, 356, 60–69.

(151) Steinmetz, N. F.; Hong, V.; Spoerke, E. D.; Lu, P.;Breitenkamp, K.; Finn, M. G.; Manchester, M. J. Am. Chem. Soc.2009, 131, 17093–17095.

(152) Douglas, S. M.; Dietz, H.; Liedl, T.; H€ogberg, B.; Graf, F.;Shih, W. M. Nature 2009, 459, 414–418.

(153) De Jonge, N.; Peckys, D. B.; Kremers, G. J.; Piston, D.W. Proc.Natl. Acad. Sci. U.S.A. 2009, 106, 2159–2164.

(154) Smith, J. E.; Wang, L.; Tan, W. Trends Anal. Chem. 2006,25, 848–855.

(155) Rayavarapu, R. G.; Petersen, W.; Ungureanu, C.; Post, J. N.; vanLeeuwen, T. G.; Manohar, S. Int. J. Biomed. Imaging 2007, 2007, 29817.

(156) Skrabalak, S. E.; Chen, J.; Sun, Y.; Lu, X.; Au, L.; Colbey,C. M.; Xia, Y. Acc. Chem. Res. 2008, 41, 1587–1595.

(157) Wang, Y.; Shen, Y.; Xie, A.; Li, S.; Wang, X.; Cai, Y. J. Phys.Chem. C 2010, 114, 4297–4301.

(158) Sastry, T. P.; Sundaraseelan, J.; Swarnalatha, K.; Liji Sobhana,S. S.; Uma Makheswari, M.; Sekar, S; Mandal, A. B. Nanotechnology2008, 19, 245604.

(159) Thiberge, S.; Zik, O.; Moses, E. Rev. Sci. Instrum. 2004,75, 2280–2289.

(160) Gatti, A. M.; Kirkpatrick, J.; Gambarelli, A.; Capitani, F.;Hansen, T.; Eloy, R.; Clermont, G. J. Mater. Sci. Mater. Med. 2008,19, 1515–1522.

(161) Gaczynska, M.; Osmulski, P. A. Curr. Opin. Colloid InterfaceSci. 2008, 13, 351–367.

(162) Cohen, S. R.; Bitler, A. Curr. Opin. Colloid Interface Sci. 2008,13, 316–325.



(163) Sitterberg, J.; Ozcetin, A.; Ehrhardt, C.; Bakowsky, U. Eur.J. Pharm. Biopharm. 2010, 74, 2–13.(164) Campbell, J. F.; Tessmer, I.; Thorp, H. H.; Erie, D. A. J. Am.

Chem. Soc. 2008, 130, 10648–10655.(165) Irrgang, J.; Ksienczyk, J.; Lapiene, V.; Niemeyer, C. M. Chem.

Phys. Chem. 2009, 10, 1483–1491.(166) Wang, H.; Tessmer, I.; Croteau, D. L.; Erie, D. A.; Van

Houten, B. Nano Lett. 2008, 8, 1631–1637.(167) Kuzuya, A.; Komiyama, M. Nanoscale 2010, 2, 310–322.(168) Rothemund, P. W. K. Nature 2006, 440, 297–302.(169) Gu, H.; Chao, J.; Xiao, S.-J.; Seeman, N. C. Nat. Nanotechnol.

2009, 4, 245–248.(170) Park, H. K.; Lim, Y. T.; Kim, J. K.; Park, H. G.; Chung, B. H.

Ultramicroscopy 2008, 108, 1115–1119.(171) Zeng, G.; Chen, J.; Zhong, L.; Wang, R.; Jiang, L.; Cai, J.; Yan,

L.; Huang, D.; Chen, C. Y.; Chen, Z. W. Proteomics 2009, 9, 1538–1547.(172) Pavia, D. L.; Lampman, G. M.; Kriz, G. S.; Vyvyan, J. A.

Introduction to Spectroscopy, 4th ed.; Brooks/Cole: Belmont, CA, 2008.(173) Biju, V.; Mundayoor, S.; Omkumar, R. V.; Anas, A.; Ishikawa,

M. Biotechnol. Adv. 2010, 28, 199–213.(174) Ohmori, S.; Saito, T.; Tange, M.; Shukla, B.; Okazaki, T.;

Yumura, M.; Iijima, S. J. Phys. Chem. C 2010, 114, 10077–10081.(175) Schwartzberg, A. M.; Zhang, J. Z. J. Phys. Chem. C 2008,

112, 10323–10337.(176) Amendola, V.; Meneghetti, M. J. Phys. Chem. C 2009,

113, 4277–4285.(177) Deka, J.; Paul, A.; Chattopadhyay, A. J. Phys. Chem. C 2009,

113, 6936–6947.(178) Hall, W. P.; Ngatia, S. N.; Van Duyne, R. P. J. Phys. Chem. C

2011, 115, 1410–1414.(179) Jiang, X.; Jiang, J.; Jin, Y.; Wang, E.; Dong, S. Biomacromole-

cules 2005, 6, 46–53.(180) Perevedentseva, E.; Cai, P.-J.; Chiu, Y.-C.; Cheng, C.-L.

Langmuir 2011, 27, 1085–1091.(181) Kandori, K.; Kuroda, T.; Togashi, S.; Katayama, E. J. Phys.

Chem. B 2011, 115, 653–659.(182) Wu, X.; Narsimhan, G. Langmuir 2008, 24, 4989–4998.(183) Olson, B. J. S. C.; Markwell, J. Curr. Protoc. Protein Sci. 2007,

3.4, 3.4.1–3.4.29.(184) Zhang, H.; Meyerhoff, M. E. Anal. Chem. 2006, 78, 609–616.(185) Ranjbar, B.; Gill, P. Chem. Biol. Drug Des. 2009, 74, 101–120.(186) Shang, L.; Wang, Y.; Jiang, J.; Dong, S. Langmuir 2007,

23, 2714–2721.(187) Asuri, P.; Bale, S. S.; Pangule, R. C.; Shah, D. A.; Kane, R. S.;

Dordick, J. S. Langmuir 2007, 23, 12318–12321.(188) Mahmoudi, M.; Shokrgozar, M. A.; Sardari, S.; Moghadam,

M. K.; Vali, H.; Laurent, S.; Stroeve, P. Nanoscale 2011, 3, 1127–1138.(189) Mamedova, N. N.; Kotov, N. A.; Rogach, A. L.; Studer, J.Nano

Lett. 2001, 1, 281–286.(190) Shanmugam, G.; Polavarapu, P. L.; Kendall, A.; Stubbs, G.

J. Gen. Virol. 2005, 86, 2371–2377.(191) Lakowicz, J. R. Principles of Fluorescence Spectroscopy, 3rd ed.;

Springer: New York, 2006.(192) Haugland, R. P. The Handbook: A Guide to Fluorescent Probes

and Labeling Technologies, 10th ed.; Invitrogen Corp.: California, 2005.(193) Nielson, L. J.; Eyley, S.; Thielemans, W.; Aylott, J. W. Chem.

Commun. 2010, 46, 8929–8931.(194) Manokaran, S.; Zhang, X.; Chen, W.; Srivastavam, D. K.

Biochim. Biophys. Acta 2010, 1804, 1376–1384.(195) Reulen, S. W. A.; Brusselaars, W. W. T.; Langereis, S.; Mulder,

W. J. M.; Breurken, M.; Merkx,M. Bioconjugate Chem. 2007, 16, 590–596.(196) Hurst, S. J.; Lytton-Jean, A. K. R.; Mirkin, C. A. Anal. Chem.

2006, 78, 8313–8318.(197) Lockney, D. M.; Guenther, R. N.; Loo, L.; Overton, W.;

Antonellu, R.; Clark, J.; Hu, M.; Luft, C.; Lommel, S. A.; Franzen, S.Bioconjugate Chem. 2011, 22, 67–73.(198) Mittal, R.; Bruchez, M. P. Bioconjugate Chem. 2011,

22, 362–368.

(199) Nakamura, M.; Ishimura, K. Langmuir 2008, 24, 12228–12234.(200) Chen, H.; Gao, J.; Lu, Y.; Kou, G.; Zhang, H.; Fan, L.; Sun, Z.;

Guo, Y.; Zhong, Y. J. Controlled Release 2008, 128, 209–216.(201) Sapsford, K. E.; Berti, L.; Medintz, I. L. Angew. Chem., Int. Ed.

2006, 45, 4562–4588.(202) Medintz, I. L.; Mattoussi, H. Phys. Chem. Chem. Phys. 2009,

11, 17–45.(203) Medintz, I. L.; Konnert, J. H.; Clapp, A. R.; Stanish, I.; Twigg,

M. E.; Mattoussi, H.; Mauro, J. M.; Deschamps, J. R. Proc. Natl. Acad. Sci.U.S.A. 2004, 101, 9612–9617.

(204) Boeneman, K.; Deschamps, J. R.; Buckhout-White, S.; Pra-suhn, D. E.; Blanco-Canosa, J. B.; Dawson, P. E.; Stewart, M. H.;Susumu, K.; Goldman, E. R.; Ancona,M.;Medintz, I. L.ACSNano 2010,4, 7253–7266.

(205) Prasuhn, D. E.; Felta, A.; Bianco-Canosa, J. B.; Susumu, K.;Stewart, M. H.; Mei, B. C.; Yakovlev, A. V.; Loukov, C.; Mallet, J. M.;Oheim,M.; Dawson, P. E.;Medintz, I. L.ACSNano 2010, 4, 5487–5497.

(206) Algar, W. R.; Tavares, A. J.; Krull, U. J. Anal. Chim. Acta 2010,673, 1–25.

(207) Morgner, F.; Geissler, D.; Stufler, S.; Butlin, N. G.;L€ohmannsr€oben, H.-G.; Hilderbrandt, N. Angew. Chem., Int. Ed. 2010,49, 7570–7574.

(208) Pons, T.; Medintz, I. L.; Sapsford, K. E.; Higashiya, S.; Grimes,A. F.; English, D. S.; Mattoussi, H. Nano Lett. 2007, 7, 3157–3164.

(209) Sen, T.; Haldar, K. K.; Patra, A. J. Phys. Chem. C 2008,112, 17945–17951.

(210) Kong, J.; Yu, S. Acta Biochem. Biophys. Sin. 2007, 38, 549–559.(211) Krpeti�c, �Z.; Nativo, P.; Porta, F.; Brust, M. A Bioconjugate

Chem. 2009, 20, 619–624.(212) Liang, Y.-Y.; Zhang, L.-M. Biomacromolecules 2007,

8, 1480–1486.(213) Susumu, K.; Uyeda, H. T.; Medintz, I. L.; Pons, T.; Delehanty,

J. B.; Mattoussi, H. J. Am. Chem. Soc. 2007, 129, 13987–13996.(214) Lundqvist, M.; Sethson, I.; Jonsson, B.-H. Biochemistry 2005,

44, 10093–10099.(215) Li, K.; Chen, Y.; Li, S.; Nguyen, G.; Niu, Z.; You, S.; Mello,

C. M.; Lu, X.; Wang, Q. Bioconjugate Chem. 2010, 21, 1369–1377.(216) Bruckman, M. A.; Kaur, G.; Lee, L. A.; Xie, F.; Sepulveda, J.;

Breitenkamp, R.; Zhang, X.; Joralemon, M.; Russell, T. P.; Emrick, T.;Wang, Q. ChemBioChem 2009, 9, 519–523.

(217) Endres, P. J.; Paunesku, T.; Vogt, S.; Meade, T. J.; Woloschak,G. E. J. Am. Chem. Soc. 2007, 129, 15760–15761.

(218) Chiang, C. K.; Chen, W. T.; Chang, H. T. Chem. Soc. Rev.2011, 40, 1269–1281.

(219) Shon, Y. S.; Choi, D.; Dare, J.; Dinh, T. Langmuir 2008,24, 6924–6931.

(220) Gibson, J. D.; Khanal, B. P.; Zubarev, E. R. J. Am. Chem. Soc.2007, 129, 11653–11661.

(221) Park, C.; Youn, H.; Kim, H.; Noh, T.; Kook, Y. H.; Oh, E. T.;Park, H. J.; Kim, C. J. Mater. Chem. 2009, 19, 2310–2315.

(222) Bothun, G. D. J. Nanobiotechnol. 2008, 6, 13.(223) Sant, S.; Thommes,M.;Hildgen, P.Langmuir 2008,24, 280–287.(224) Liu, J.; Gong, T.; Wang, C.; Zhong, Z.; Zhang, Z. Int. J. Pharm.

2007, 340, 153–162.(225) Yuan, W.; Wu, F.; Geng, Y.; Xu, S.; Jin, T. Nanotechnology

2007, 18, 145601.(226) Cedervall, T.; Lynch, I.; Berggard, T.; Thulin, E.; Nilsson, H.;

Dawson, K. A.; Linse, S. Proc. Natl. Acad. Sci. U.S.A. 2007,104, 2050–2055.

(227) Duhr, S.; Braun, D. Proc. Natl. Acad. Sci. U.S.A. 2006,103, 19678–19682.

(228) Piazza, R.; Parola, A J. Phys.: Condens. Matter 2008,20, 153102.

(229) Blundell, T. L.; Sibanda, B. L.; Sternberg, M. J.; Thornton,J. M. Nature 1987, 326, 347–352.

(230) Greer, J. Proteins 1990, 7, 317–334.(231) Sali, A.; Blundell, T. L. J. Mol. Biol. 1993, 234, 779–815.(232) Sanchez, R.; Sali, A. Methods Mol. Biol. 2000, 143, 97–129.



(233) Chothia, C.; Lesk, A. M. EMBO J. 1986, 5, 823–826.(234) Saunders, S. R.; Eden, M. R.; Roberts, C. B. J. Phys. Chem. C

2011, 115, 4603–4610.(235) Kim, Y.; Klutz, A. M.; Hechler, B.; Gao, Z.-G.; Gachet, C.;

Jacobson, K. A. Purinergic Signalling 2009, 5, 39–50.(236) Mukherjee, A. K.; Kumar, T. S.; Rai, S. K.; Roy, J. K. Biotechnol.

Bioprocess Eng. 2010, 15, 984–992.(237) Medintz, I. L.; Deschamps, J. R. Curr. Opin. Biotechnol. 2006,

17, 17–27.(238) Prasuhn, D. E.; Deschamps, J. R.; Susumu, K.; Boeneman, K.;

Blanco-Canosa, J. B.; Dawson, P. E.; Medintz, I. L. Small 2010,6, 555–564.(239) Boeneman, K.; Mei, B. C.; Dennis, A.M.; Bao, G.; Deschamps,

J. R.; Mattoussi, H.; Medintz, I. L. J. Am. Chem. Soc. 2009,131, 3828–3829.(240) Hutter, E.; Boridy, S.; Labrecque, S.; Lalancette-Hebert, M.;

Kriz, J.; Winnik, F. M.; Maysinger, D. ACS Nano 2010, 4, 2595–2606.(241) Fairhurst, D. http://www.xigonanotools.com/applications/

applications.asp (accessed March 24, 2011).(242) Ferreira, G. N. M.; da-Silva, A.-C.; Tom�e, B. Trends Biotechnol.

2009, 27, 689–697.(243) Cheng, Z.-G.; Tang, D.-Y. Bioprocess Biosyst. Eng. 2007,

30, 243–249.(244) Amstad, E.; Zurcher, S.; Mashagi, A.; Wong, J. Y.; Textor, M.;

Reimhult, E. Small 2009, 5, 1334–1342.(245) Burg, T. P.; Godin, M.; Knudsen, S. M.; Shen,W.; Carlson, G.;

Foster, J. S.; Babcock, K.; Manalis, S. R. Nature 2007, 446, 1066–1069.(246) Montes-Burgos, I.; Walczyk, D.; Hole, P.; Smith, J.; Lynch, I.;

Dawson, K. J. Nanopart. Res. 2010, 12, 47–53.(247) Braeckmans, K.; Buyens, K.; Bouquet, W.; Vervaet, C.; Joye,

P.; De Vos, F.; Plawinski, L.; Doeuvre, L.; Angles-Cano, E.; Sanders,N. N.; Demeester, J.; De Smedt, S. C. Nano Lett. 2010, 10, 4435–4442.(248) Fraikin, J.-L.; Teesalu, T.; McKenney, C. M.; Ruoslahti, E.;

Cleland, A. N. Nat. Nanotechnol. 2011, DOI: 10.1038/nnano.2011.24.(249) Willmott, G. R.; Vogel, R.; Yu, S. S. C.; Groenewegen, L. G.;

Roberts, G. S.; Kozak, D.; Anderson, W.; Trau, M. J. Phys.: Condens.Matter. 2010, 22, 454116.(250) Roberts, G. S.; Kozak, D.; Anderson,W.; Broom,M. F.; Vogel,

R.; Trau, M. Small 2010, 6, 2653–2658.(251) Monopoli, M. P.; Walczyk, D.; Campbell, A.; Elia, G.; Lynch, I.;

Bombelli, F. B.; Dawson, K. A. J. Am. Chem. Soc. 2011, 133, 2525–2534.(252) Marnett, L. J. Chem. Res. Toxicol. 2009, 22, 1491.(253) Sahu, S. C.; Casciano, D. A. Nanotoxicity: From in Vivo and in

Vitro Models to Health Risks; Wiley: New York, 2009.(254) Landsiedel, R.; Kapp, M. D.; Schulz, M.; Wiench, K.; Oesch, F.

Mutat. Res. 2009, 681, 241–258.(255) Monteiro-Riviere, N. A.; Inman, A. O.; Zhang, L. W. Toxicol.

Appl. Pharmacol. 2009, 234, 222–235.(256) Dobrovolskaia, M. A.; Germolec, D. R.; Waever, J. L. Nat.

Nanotechnol. 2009, 4, 411–414.(257) Franca, A.; Pelaz, B.; Moros, M.; Sanchez-Espinel, C.; Hernan-

dez, A.; Fernandez-Lopez, C.; Grazu, V.; de la Fuente, J.; Pastoriza-Santos,I.; Liz-Marzan, L. M.; Gonzalez-Fernandez, A. Small 2010, 6, 89–95.(258) Bellucci, S.; Chiaretti, M.; Onorato, P.; Rossella, F.; Grandi,

M. S.; Galinetto, P.; Sacco, I.; Micciulla, F. Nanomedicine 2010,5, 209–215.(259) Ianaro, A.; Tersigni, M.; D’Acquisto, F. Mini-Rev. Med. Chem.

2009, 9, 306–317.(260) Vallhov, H.; Qin, J.; Johansson, S. M.; Ahlorg, N.; Muhammed,

M. A.; Scheynius, A.; Gabrielsson, S. Nano Lett. 2006, 6, 1382–1686.(261) Esch, R. K.; Han, L.; Foarde, K. K.; Ensor, D. S.Nanotoxicology

2010, 4, 73–83.(262) Dobrovolskaia, M. A.; Neun, B. W.; Clogston, J. D.; Ding, H.;

Ljubimova, J.; McNeil, S. E. Nanomedicine 2010, 5, 555–562.

ac200853aanalyzing nano material bio conjugates a review of current and emerging purification and...

Documents

oce of science

materials science

nmbiological hybrid

unique nm property

biological hybrids

biological setting

biomolecular science

division of chemistry